In [1]:
import pandas as pd
data = pd.read_csv('https://cursopypagina.github.io/CursoPy/milknew.csv')
data
Out[1]:
pH Temprature Taste Odor Fat Turbidity Colour Grade
0 6.6 35 1 0 1 0 254 high
1 6.6 36 0 1 0 1 253 high
2 8.5 70 1 1 1 1 246 low
3 9.5 34 1 1 0 1 255 low
4 6.6 37 0 0 0 0 255 medium
... ... ... ... ... ... ... ... ...
1054 6.7 45 1 1 0 0 247 medium
1055 6.7 38 1 0 1 0 255 high
1056 3.0 40 1 1 1 1 255 low
1057 6.8 43 1 0 1 0 250 high
1058 8.6 55 0 1 1 1 255 low

1059 rows × 8 columns

In [2]:
# Importaciones necesarias
from sklearn.model_selection import train_test_split
from sklearn.linear_model import LogisticRegression
from sklearn.metrics import confusion_matrix, ConfusionMatrixDisplay
from sklearn.metrics import f1_score, recall_score, precision_score, accuracy_score
from sklearn.preprocessing import OneHotEncoder
from keras.models import Sequential
from keras.layers.core import Activation
from keras.layers import Dense
import tensorflow as tf 
import matplotlib.pyplot as plt
import matplotlib as mpl
import pandas as pd
import numpy as np
import warnings

# configuracion para no ver en salida las advertencias
warnings.simplefilter("ignore")

# Configuramos el estilo de graficacion
plt.style.use('seaborn')

# Instanciamos y configuramos one hot encoder
enc = OneHotEncoder(handle_unknown='ignore')
In [3]:
data.head()
Out[3]:
pH Temprature Taste Odor Fat Turbidity Colour Grade
0 6.6 35 1 0 1 0 254 high
1 6.6 36 0 1 0 1 253 high
2 8.5 70 1 1 1 1 246 low
3 9.5 34 1 1 0 1 255 low
4 6.6 37 0 0 0 0 255 medium
In [7]:
# Definicion de las variables
X = data.drop('Grade', axis=1).values
y = data.Grade.values.reshape(-1,1)

# Conjuntos de entrenamiento y prueba
X_train, X_test, y_train, y_test = train_test_split(X, y, 
                                                    test_size = 0.3, 
                                                    shuffle = True, 
                                                    random_state = 123)
In [8]:
# Aplicamos el one hot encoder a y_train y y_test
y_train = enc.fit_transform(y_train).toarray()
y_test = enc.fit_transform(y_test).toarray()
In [10]:
# Definimos una funcion para entrenar la red neuronal
# con distinto numero de neuronas, epocas y batch size
def train_eval_model(x1, y1, x2, y2, n1, n2, n, k):
    """Función para realizar el entrenamiento de la red neuronal. Asimismo,
    se realiza la evaluación del mismo.
    Parámetros:
    * x1: X_train.
    * y1: y_train.
    * x2: X_test.
    * y2: y_test.
    * n1: número de neuronas en la primer capa oculta.
    * n2: número de neuronas en la segunda capa oculta.
    * n: número de épocas.
    * k: batch_size."""
    # Instanciamos el modelo secuencial
    model = Sequential()
    # Capa de entrada y primera capa oculta 
    model.add(Dense(n1, activation='relu', input_shape=(7,)))
    # Capa oculta 2
    model.add(Dense(n2, activation='relu'))
    # Capa de salida
    model.add(Dense(3, activation='softmax'))
    # Copilacion del modelo
    model.compile(loss='categorical_crossentropy', optimizer='adam', metrics=['accuracy'])
    # Entrenamiento del modelo.
    # Destinamos el 15% de los datos de entrenamiento para la validacion 
    history = model.fit(x1, y1, epochs=n, validation_split=0.15, batch_size=k)
    # Evalucacion del modelo
    scores = model.evaluate(x2, y2)
    print('%s: %.4f%%' % (model.metrics_names[1], scores[1] * 100))
    # Grafico de la precision del entrenamiento vs precision de la validacion
    plt.plot(history.history['accuracy'])
    plt.plot(history.history['val_accuracy'])
    plt.title('Precisión del modelo')
    plt.xlabel('Epoch')
    plt.ylabel('Precisión')
    plt.legend(['Train', 'Validación'])
    plt.show()
In [11]:
train_eval_model(X_train, y_train, X_test, y_test, 56, 21, 5, 5)

Epoch 1/5
126/126 [==============================] - 2s 4ms/step - loss: 3.1685 - accuracy: 0.3386 - val_loss: 1.4411 - val_accuracy: 0.4018
Epoch 2/5
126/126 [==============================] - 0s 2ms/step - loss: 1.2068 - accuracy: 0.4420 - val_loss: 1.4002 - val_accuracy: 0.4643
Epoch 3/5
126/126 [==============================] - 0s 2ms/step - loss: 1.0785 - accuracy: 0.4722 - val_loss: 1.0307 - val_accuracy: 0.3750
Epoch 4/5
126/126 [==============================] - 0s 2ms/step - loss: 1.0600 - accuracy: 0.5087 - val_loss: 1.0377 - val_accuracy: 0.5446
Epoch 5/5
126/126 [==============================] - 0s 2ms/step - loss: 1.0842 - accuracy: 0.5167 - val_loss: 1.4827 - val_accuracy: 0.4464
10/10 [==============================] - 0s 1ms/step - loss: 1.3713 - accuracy: 0.4969
accuracy: 49.6855%
In [12]:
train_eval_model(X_train, y_train, X_test, y_test, 56, 21, 5, 10)

Epoch 1/5
63/63 [==============================] - 1s 4ms/step - loss: 4.7327 - accuracy: 0.4213 - val_loss: 0.9677 - val_accuracy: 0.5625
Epoch 2/5
63/63 [==============================] - 0s 2ms/step - loss: 1.0545 - accuracy: 0.4992 - val_loss: 1.0372 - val_accuracy: 0.4375
Epoch 3/5
63/63 [==============================] - 0s 2ms/step - loss: 0.9547 - accuracy: 0.5358 - val_loss: 1.0120 - val_accuracy: 0.5089
Epoch 4/5
63/63 [==============================] - 0s 2ms/step - loss: 0.9430 - accuracy: 0.5374 - val_loss: 1.0082 - val_accuracy: 0.5179
Epoch 5/5
63/63 [==============================] - 0s 2ms/step - loss: 0.9527 - accuracy: 0.5564 - val_loss: 1.0125 - val_accuracy: 0.4821
10/10 [==============================] - 0s 1ms/step - loss: 0.9320 - accuracy: 0.5755
accuracy: 57.5472%
In [13]:
train_eval_model(X_train, y_train, X_test, y_test, 56, 21, 15, 10)

Epoch 1/15
63/63 [==============================] - 1s 4ms/step - loss: 3.7845 - accuracy: 0.3577 - val_loss: 1.3080 - val_accuracy: 0.3304
Epoch 2/15
63/63 [==============================] - 0s 2ms/step - loss: 1.0080 - accuracy: 0.4897 - val_loss: 0.8929 - val_accuracy: 0.5714
Epoch 3/15
63/63 [==============================] - 0s 2ms/step - loss: 0.8928 - accuracy: 0.5739 - val_loss: 1.2841 - val_accuracy: 0.5089
Epoch 4/15
63/63 [==============================] - 0s 2ms/step - loss: 0.9794 - accuracy: 0.5930 - val_loss: 1.0313 - val_accuracy: 0.4375
Epoch 5/15
63/63 [==============================] - 0s 2ms/step - loss: 0.9300 - accuracy: 0.5501 - val_loss: 0.9576 - val_accuracy: 0.5089
Epoch 6/15
63/63 [==============================] - 0s 2ms/step - loss: 0.9003 - accuracy: 0.5914 - val_loss: 0.8889 - val_accuracy: 0.5357
Epoch 7/15
63/63 [==============================] - 0s 2ms/step - loss: 0.8047 - accuracy: 0.6391 - val_loss: 0.9561 - val_accuracy: 0.5714
Epoch 8/15
63/63 [==============================] - 0s 2ms/step - loss: 0.8513 - accuracy: 0.6328 - val_loss: 0.8716 - val_accuracy: 0.6429
Epoch 9/15
63/63 [==============================] - 0s 2ms/step - loss: 0.9319 - accuracy: 0.5946 - val_loss: 1.2490 - val_accuracy: 0.4732
Epoch 10/15
63/63 [==============================] - 0s 2ms/step - loss: 0.9445 - accuracy: 0.6105 - val_loss: 0.7887 - val_accuracy: 0.5714
Epoch 11/15
63/63 [==============================] - 0s 2ms/step - loss: 0.8398 - accuracy: 0.6264 - val_loss: 0.7716 - val_accuracy: 0.6339
Epoch 12/15
63/63 [==============================] - 0s 2ms/step - loss: 0.7838 - accuracy: 0.6439 - val_loss: 0.7622 - val_accuracy: 0.6250
Epoch 13/15
63/63 [==============================] - 0s 2ms/step - loss: 0.7908 - accuracy: 0.6391 - val_loss: 0.7817 - val_accuracy: 0.6696
Epoch 14/15
63/63 [==============================] - 0s 2ms/step - loss: 0.7179 - accuracy: 0.6741 - val_loss: 0.7689 - val_accuracy: 0.5804
Epoch 15/15
63/63 [==============================] - 0s 2ms/step - loss: 0.7957 - accuracy: 0.6455 - val_loss: 0.7369 - val_accuracy: 0.5982
10/10 [==============================] - 0s 1ms/step - loss: 0.6330 - accuracy: 0.6667
accuracy: 66.6667%
In [14]:
train_eval_model(X_train, y_train, X_test, y_test, 70, 40, 15, 10)

Epoch 1/15
63/63 [==============================] - 1s 5ms/step - loss: 2.6239 - accuracy: 0.3625 - val_loss: 1.0043 - val_accuracy: 0.4732
Epoch 2/15
63/63 [==============================] - 0s 2ms/step - loss: 1.2542 - accuracy: 0.4404 - val_loss: 1.1436 - val_accuracy: 0.5089
Epoch 3/15
63/63 [==============================] - 0s 2ms/step - loss: 1.0701 - accuracy: 0.5008 - val_loss: 0.9438 - val_accuracy: 0.5179
Epoch 4/15
63/63 [==============================] - 0s 2ms/step - loss: 1.0312 - accuracy: 0.5151 - val_loss: 1.2310 - val_accuracy: 0.5357
Epoch 5/15
63/63 [==============================] - 0s 2ms/step - loss: 1.1047 - accuracy: 0.5533 - val_loss: 1.6473 - val_accuracy: 0.5089
Epoch 6/15
63/63 [==============================] - 0s 2ms/step - loss: 1.2741 - accuracy: 0.4976 - val_loss: 1.3502 - val_accuracy: 0.4821
Epoch 7/15
63/63 [==============================] - 0s 2ms/step - loss: 1.1664 - accuracy: 0.5469 - val_loss: 1.6382 - val_accuracy: 0.5179
Epoch 8/15
63/63 [==============================] - 0s 2ms/step - loss: 0.9417 - accuracy: 0.6169 - val_loss: 1.0166 - val_accuracy: 0.5714
Epoch 9/15
63/63 [==============================] - 0s 2ms/step - loss: 1.5314 - accuracy: 0.5374 - val_loss: 1.3891 - val_accuracy: 0.5536
Epoch 10/15
63/63 [==============================] - 0s 2ms/step - loss: 0.9487 - accuracy: 0.5914 - val_loss: 1.3912 - val_accuracy: 0.5536
Epoch 11/15
63/63 [==============================] - 0s 2ms/step - loss: 1.2958 - accuracy: 0.5485 - val_loss: 0.8340 - val_accuracy: 0.5982
Epoch 12/15
63/63 [==============================] - 0s 2ms/step - loss: 0.7369 - accuracy: 0.6407 - val_loss: 0.7104 - val_accuracy: 0.5714
Epoch 13/15
63/63 [==============================] - 0s 2ms/step - loss: 0.7699 - accuracy: 0.6439 - val_loss: 0.9825 - val_accuracy: 0.6250
Epoch 14/15
63/63 [==============================] - 0s 2ms/step - loss: 1.0924 - accuracy: 0.5819 - val_loss: 1.0766 - val_accuracy: 0.5000
Epoch 15/15
63/63 [==============================] - 0s 2ms/step - loss: 0.9561 - accuracy: 0.6121 - val_loss: 0.7275 - val_accuracy: 0.6250
10/10 [==============================] - 0s 1ms/step - loss: 0.6356 - accuracy: 0.7075
accuracy: 70.7547%
In [18]:
train_eval_model(X_train, y_train, X_test, y_test, 70, 20, 15, 10)

Epoch 1/15
63/63 [==============================] - 1s 4ms/step - loss: 2.6935 - accuracy: 0.4372 - val_loss: 0.9719 - val_accuracy: 0.4643
Epoch 2/15
63/63 [==============================] - 0s 2ms/step - loss: 1.0157 - accuracy: 0.4928 - val_loss: 0.9554 - val_accuracy: 0.5536
Epoch 3/15
63/63 [==============================] - 0s 2ms/step - loss: 0.9899 - accuracy: 0.5278 - val_loss: 1.0383 - val_accuracy: 0.5982
Epoch 4/15
63/63 [==============================] - 0s 2ms/step - loss: 0.9386 - accuracy: 0.5533 - val_loss: 1.1095 - val_accuracy: 0.4911
Epoch 5/15
63/63 [==============================] - 0s 2ms/step - loss: 1.0454 - accuracy: 0.5215 - val_loss: 0.8732 - val_accuracy: 0.6250
Epoch 6/15
63/63 [==============================] - 0s 2ms/step - loss: 0.8858 - accuracy: 0.5771 - val_loss: 0.8798 - val_accuracy: 0.5714
Epoch 7/15
63/63 [==============================] - 0s 2ms/step - loss: 0.8394 - accuracy: 0.5994 - val_loss: 0.8703 - val_accuracy: 0.7232
Epoch 8/15
63/63 [==============================] - 0s 2ms/step - loss: 0.8251 - accuracy: 0.6137 - val_loss: 0.8250 - val_accuracy: 0.5625
Epoch 9/15
63/63 [==============================] - 0s 2ms/step - loss: 0.9355 - accuracy: 0.5755 - val_loss: 0.8456 - val_accuracy: 0.6250
Epoch 10/15
63/63 [==============================] - 0s 2ms/step - loss: 0.7967 - accuracy: 0.6312 - val_loss: 0.8838 - val_accuracy: 0.6071
Epoch 11/15
63/63 [==============================] - 0s 2ms/step - loss: 0.8424 - accuracy: 0.6407 - val_loss: 0.9869 - val_accuracy: 0.5000
Epoch 12/15
63/63 [==============================] - 0s 2ms/step - loss: 0.8194 - accuracy: 0.6137 - val_loss: 0.7939 - val_accuracy: 0.5536
Epoch 13/15
63/63 [==============================] - 0s 2ms/step - loss: 0.7914 - accuracy: 0.6439 - val_loss: 0.9326 - val_accuracy: 0.5536
Epoch 14/15
63/63 [==============================] - 0s 2ms/step - loss: 0.8101 - accuracy: 0.6264 - val_loss: 0.7857 - val_accuracy: 0.6607
Epoch 15/15
63/63 [==============================] - 0s 2ms/step - loss: 0.7860 - accuracy: 0.6423 - val_loss: 0.8437 - val_accuracy: 0.6518
10/10 [==============================] - 0s 2ms/step - loss: 0.7449 - accuracy: 0.7044
accuracy: 70.4403%
In [19]:
train_eval_model(X_train, y_train, X_test, y_test, 70, 20, 15, 20)

Epoch 1/15
32/32 [==============================] - 1s 7ms/step - loss: 2.2521 - accuracy: 0.4149 - val_loss: 1.1874 - val_accuracy: 0.4911
Epoch 2/15
32/32 [==============================] - 0s 2ms/step - loss: 0.9733 - accuracy: 0.4992 - val_loss: 1.0319 - val_accuracy: 0.5179
Epoch 3/15
32/32 [==============================] - 0s 2ms/step - loss: 0.9897 - accuracy: 0.4849 - val_loss: 0.9562 - val_accuracy: 0.5893
Epoch 4/15
32/32 [==============================] - 0s 2ms/step - loss: 0.8979 - accuracy: 0.5580 - val_loss: 1.3028 - val_accuracy: 0.4286
Epoch 5/15
32/32 [==============================] - 0s 3ms/step - loss: 0.9536 - accuracy: 0.5517 - val_loss: 1.1055 - val_accuracy: 0.4554
Epoch 6/15
32/32 [==============================] - 0s 2ms/step - loss: 0.8848 - accuracy: 0.5501 - val_loss: 0.9072 - val_accuracy: 0.6161
Epoch 7/15
32/32 [==============================] - 0s 2ms/step - loss: 0.8419 - accuracy: 0.6041 - val_loss: 0.9487 - val_accuracy: 0.5446
Epoch 8/15
32/32 [==============================] - 0s 2ms/step - loss: 0.9651 - accuracy: 0.5517 - val_loss: 0.9849 - val_accuracy: 0.5268
Epoch 9/15
32/32 [==============================] - 0s 2ms/step - loss: 0.8533 - accuracy: 0.6041 - val_loss: 0.9385 - val_accuracy: 0.5357
Epoch 10/15
32/32 [==============================] - 0s 2ms/step - loss: 0.9021 - accuracy: 0.5707 - val_loss: 0.9890 - val_accuracy: 0.5536
Epoch 11/15
32/32 [==============================] - 0s 2ms/step - loss: 0.8762 - accuracy: 0.5755 - val_loss: 0.9371 - val_accuracy: 0.5893
Epoch 12/15
32/32 [==============================] - 0s 2ms/step - loss: 0.8078 - accuracy: 0.6566 - val_loss: 0.8728 - val_accuracy: 0.5982
Epoch 13/15
32/32 [==============================] - 0s 2ms/step - loss: 0.8271 - accuracy: 0.6105 - val_loss: 1.0003 - val_accuracy: 0.5804
Epoch 14/15
32/32 [==============================] - 0s 2ms/step - loss: 0.8794 - accuracy: 0.5978 - val_loss: 0.9035 - val_accuracy: 0.6696
Epoch 15/15
32/32 [==============================] - 0s 2ms/step - loss: 0.7825 - accuracy: 0.6375 - val_loss: 0.9728 - val_accuracy: 0.5714
10/10 [==============================] - 0s 1ms/step - loss: 0.8282 - accuracy: 0.5975
accuracy: 59.7484%
In [20]:
train_eval_model(X_train, y_train, X_test, y_test, 70, 20, 50, 10)

Epoch 1/50
63/63 [==============================] - 1s 11ms/step - loss: 2.1733 - accuracy: 0.4881 - val_loss: 1.0057 - val_accuracy: 0.4911
Epoch 2/50
63/63 [==============================] - 0s 2ms/step - loss: 1.1373 - accuracy: 0.4515 - val_loss: 1.4169 - val_accuracy: 0.4464
Epoch 3/50
63/63 [==============================] - 0s 2ms/step - loss: 1.0717 - accuracy: 0.5151 - val_loss: 1.1205 - val_accuracy: 0.5089
Epoch 4/50
63/63 [==============================] - 0s 2ms/step - loss: 0.9602 - accuracy: 0.5437 - val_loss: 1.0729 - val_accuracy: 0.5268
Epoch 5/50
63/63 [==============================] - 0s 2ms/step - loss: 0.9187 - accuracy: 0.5723 - val_loss: 1.1219 - val_accuracy: 0.5357
Epoch 6/50
63/63 [==============================] - 0s 2ms/step - loss: 1.0370 - accuracy: 0.5787 - val_loss: 1.0303 - val_accuracy: 0.4821
Epoch 7/50
63/63 [==============================] - 0s 2ms/step - loss: 0.8515 - accuracy: 0.6105 - val_loss: 0.7796 - val_accuracy: 0.6339
Epoch 8/50
63/63 [==============================] - 0s 2ms/step - loss: 0.8056 - accuracy: 0.6359 - val_loss: 1.0217 - val_accuracy: 0.4286
Epoch 9/50
63/63 [==============================] - 0s 2ms/step - loss: 0.8857 - accuracy: 0.5946 - val_loss: 0.8369 - val_accuracy: 0.5446
Epoch 10/50
63/63 [==============================] - 0s 2ms/step - loss: 0.7470 - accuracy: 0.6534 - val_loss: 0.7234 - val_accuracy: 0.6339
Epoch 11/50
63/63 [==============================] - 0s 2ms/step - loss: 0.7177 - accuracy: 0.6550 - val_loss: 0.6930 - val_accuracy: 0.7054
Epoch 12/50
63/63 [==============================] - 0s 2ms/step - loss: 0.7718 - accuracy: 0.6614 - val_loss: 0.6827 - val_accuracy: 0.6607
Epoch 13/50
63/63 [==============================] - 0s 2ms/step - loss: 0.7125 - accuracy: 0.6709 - val_loss: 0.6563 - val_accuracy: 0.7321
Epoch 14/50
63/63 [==============================] - 0s 2ms/step - loss: 0.6580 - accuracy: 0.6995 - val_loss: 1.0750 - val_accuracy: 0.4732
Epoch 15/50
63/63 [==============================] - 0s 2ms/step - loss: 0.7156 - accuracy: 0.6804 - val_loss: 0.6423 - val_accuracy: 0.6786
Epoch 16/50
63/63 [==============================] - 0s 2ms/step - loss: 0.7135 - accuracy: 0.6836 - val_loss: 0.7776 - val_accuracy: 0.6518
Epoch 17/50
63/63 [==============================] - 0s 2ms/step - loss: 0.7664 - accuracy: 0.6471 - val_loss: 0.7031 - val_accuracy: 0.6161
Epoch 18/50
63/63 [==============================] - 0s 2ms/step - loss: 0.6253 - accuracy: 0.7154 - val_loss: 0.8211 - val_accuracy: 0.6161
Epoch 19/50
63/63 [==============================] - 0s 2ms/step - loss: 0.6740 - accuracy: 0.7059 - val_loss: 0.9292 - val_accuracy: 0.5536
Epoch 20/50
63/63 [==============================] - 0s 2ms/step - loss: 0.6304 - accuracy: 0.7122 - val_loss: 0.6291 - val_accuracy: 0.7411
Epoch 21/50
63/63 [==============================] - 0s 2ms/step - loss: 0.6111 - accuracy: 0.7329 - val_loss: 1.0594 - val_accuracy: 0.4732
Epoch 22/50
63/63 [==============================] - 0s 2ms/step - loss: 0.6608 - accuracy: 0.7059 - val_loss: 0.6953 - val_accuracy: 0.6429
Epoch 23/50
63/63 [==============================] - 0s 2ms/step - loss: 0.6052 - accuracy: 0.7393 - val_loss: 0.8263 - val_accuracy: 0.6071
Epoch 24/50
63/63 [==============================] - 0s 2ms/step - loss: 0.5594 - accuracy: 0.7583 - val_loss: 0.6065 - val_accuracy: 0.7321
Epoch 25/50
63/63 [==============================] - 0s 2ms/step - loss: 0.6282 - accuracy: 0.7059 - val_loss: 0.5127 - val_accuracy: 0.7946
Epoch 26/50
63/63 [==============================] - 0s 2ms/step - loss: 0.5634 - accuracy: 0.7456 - val_loss: 0.6908 - val_accuracy: 0.6875
Epoch 27/50
63/63 [==============================] - 0s 2ms/step - loss: 0.4948 - accuracy: 0.7758 - val_loss: 0.7551 - val_accuracy: 0.7500
Epoch 28/50
63/63 [==============================] - 0s 2ms/step - loss: 0.6791 - accuracy: 0.7027 - val_loss: 0.6721 - val_accuracy: 0.7500
Epoch 29/50
63/63 [==============================] - 0s 2ms/step - loss: 0.4952 - accuracy: 0.7917 - val_loss: 0.4764 - val_accuracy: 0.7500
Epoch 30/50
63/63 [==============================] - 0s 2ms/step - loss: 0.6671 - accuracy: 0.7250 - val_loss: 0.5425 - val_accuracy: 0.8125
Epoch 31/50
63/63 [==============================] - 0s 2ms/step - loss: 0.4858 - accuracy: 0.7886 - val_loss: 0.5404 - val_accuracy: 0.7500
Epoch 32/50
63/63 [==============================] - 0s 2ms/step - loss: 0.4597 - accuracy: 0.8076 - val_loss: 0.5148 - val_accuracy: 0.8036
Epoch 33/50
63/63 [==============================] - 0s 2ms/step - loss: 0.4940 - accuracy: 0.7997 - val_loss: 0.4788 - val_accuracy: 0.7768
Epoch 34/50
63/63 [==============================] - 0s 2ms/step - loss: 0.5209 - accuracy: 0.7631 - val_loss: 0.5457 - val_accuracy: 0.7321
Epoch 35/50
63/63 [==============================] - 0s 2ms/step - loss: 0.5848 - accuracy: 0.7456 - val_loss: 0.4299 - val_accuracy: 0.7946
Epoch 36/50
63/63 [==============================] - 0s 2ms/step - loss: 0.4850 - accuracy: 0.7774 - val_loss: 0.7103 - val_accuracy: 0.6696
Epoch 37/50
63/63 [==============================] - 0s 2ms/step - loss: 0.6031 - accuracy: 0.7329 - val_loss: 0.5480 - val_accuracy: 0.6786
Epoch 38/50
63/63 [==============================] - 0s 2ms/step - loss: 0.5247 - accuracy: 0.7854 - val_loss: 0.4151 - val_accuracy: 0.8393
Epoch 39/50
63/63 [==============================] - 0s 2ms/step - loss: 0.4943 - accuracy: 0.7854 - val_loss: 0.7193 - val_accuracy: 0.7232
Epoch 40/50
63/63 [==============================] - 0s 2ms/step - loss: 0.4043 - accuracy: 0.8156 - val_loss: 0.4433 - val_accuracy: 0.7321
Epoch 41/50
63/63 [==============================] - 0s 2ms/step - loss: 0.3982 - accuracy: 0.8347 - val_loss: 0.4272 - val_accuracy: 0.8304
Epoch 42/50
63/63 [==============================] - 0s 2ms/step - loss: 0.3788 - accuracy: 0.8442 - val_loss: 0.4213 - val_accuracy: 0.8125
Epoch 43/50
63/63 [==============================] - 0s 2ms/step - loss: 0.5400 - accuracy: 0.7663 - val_loss: 0.4300 - val_accuracy: 0.8304
Epoch 44/50
63/63 [==============================] - 0s 2ms/step - loss: 0.5264 - accuracy: 0.7806 - val_loss: 0.4973 - val_accuracy: 0.8125
Epoch 45/50
63/63 [==============================] - 0s 2ms/step - loss: 0.5108 - accuracy: 0.7742 - val_loss: 0.8640 - val_accuracy: 0.6518
Epoch 46/50
63/63 [==============================] - 0s 2ms/step - loss: 0.5019 - accuracy: 0.7933 - val_loss: 0.4425 - val_accuracy: 0.8393
Epoch 47/50
63/63 [==============================] - 0s 2ms/step - loss: 0.4211 - accuracy: 0.8203 - val_loss: 0.3530 - val_accuracy: 0.8839
Epoch 48/50
63/63 [==============================] - 0s 2ms/step - loss: 0.3360 - accuracy: 0.8521 - val_loss: 0.3794 - val_accuracy: 0.8393
Epoch 49/50
63/63 [==============================] - 0s 2ms/step - loss: 0.3938 - accuracy: 0.8378 - val_loss: 0.3881 - val_accuracy: 0.8125
Epoch 50/50
63/63 [==============================] - 0s 2ms/step - loss: 0.4570 - accuracy: 0.8108 - val_loss: 0.4446 - val_accuracy: 0.8036
10/10 [==============================] - 0s 1ms/step - loss: 0.3953 - accuracy: 0.8050
accuracy: 80.5031%
In [21]:
train_eval_model(X_train, y_train, X_test, y_test, 70, 30, 50, 10)

Epoch 1/50
63/63 [==============================] - 1s 4ms/step - loss: 1.5255 - accuracy: 0.4833 - val_loss: 0.9487 - val_accuracy: 0.4643
Epoch 2/50
63/63 [==============================] - 0s 2ms/step - loss: 0.9918 - accuracy: 0.5342 - val_loss: 1.1076 - val_accuracy: 0.4018
Epoch 3/50
63/63 [==============================] - 0s 2ms/step - loss: 0.9721 - accuracy: 0.5167 - val_loss: 1.4354 - val_accuracy: 0.4286
Epoch 4/50
63/63 [==============================] - 0s 2ms/step - loss: 0.9738 - accuracy: 0.5612 - val_loss: 0.9609 - val_accuracy: 0.5089
Epoch 5/50
63/63 [==============================] - 0s 2ms/step - loss: 1.1748 - accuracy: 0.5024 - val_loss: 1.5074 - val_accuracy: 0.4286
Epoch 6/50
63/63 [==============================] - 0s 2ms/step - loss: 0.9197 - accuracy: 0.5707 - val_loss: 0.9434 - val_accuracy: 0.4464
Epoch 7/50
63/63 [==============================] - 0s 2ms/step - loss: 0.8445 - accuracy: 0.5978 - val_loss: 0.9208 - val_accuracy: 0.5268
Epoch 8/50
63/63 [==============================] - 0s 2ms/step - loss: 0.8219 - accuracy: 0.5962 - val_loss: 1.0765 - val_accuracy: 0.4821
Epoch 9/50
63/63 [==============================] - 0s 2ms/step - loss: 0.7632 - accuracy: 0.6391 - val_loss: 1.2100 - val_accuracy: 0.5268
Epoch 10/50
63/63 [==============================] - 0s 2ms/step - loss: 0.8357 - accuracy: 0.6216 - val_loss: 0.8581 - val_accuracy: 0.6964
Epoch 11/50
63/63 [==============================] - 0s 2ms/step - loss: 0.8154 - accuracy: 0.6375 - val_loss: 0.8581 - val_accuracy: 0.5714
Epoch 12/50
63/63 [==============================] - 0s 2ms/step - loss: 0.7586 - accuracy: 0.6407 - val_loss: 0.8728 - val_accuracy: 0.5714
Epoch 13/50
63/63 [==============================] - 0s 2ms/step - loss: 0.8085 - accuracy: 0.6423 - val_loss: 0.9593 - val_accuracy: 0.6161
Epoch 14/50
63/63 [==============================] - 0s 2ms/step - loss: 0.7932 - accuracy: 0.6105 - val_loss: 1.1950 - val_accuracy: 0.4732
Epoch 15/50
63/63 [==============================] - 0s 2ms/step - loss: 0.9688 - accuracy: 0.6200 - val_loss: 0.8666 - val_accuracy: 0.6250
Epoch 16/50
63/63 [==============================] - 0s 2ms/step - loss: 0.8832 - accuracy: 0.6010 - val_loss: 0.6854 - val_accuracy: 0.6429
Epoch 17/50
63/63 [==============================] - 0s 2ms/step - loss: 0.7254 - accuracy: 0.6868 - val_loss: 0.6608 - val_accuracy: 0.6339
Epoch 18/50
63/63 [==============================] - 0s 2ms/step - loss: 0.7126 - accuracy: 0.6534 - val_loss: 0.6865 - val_accuracy: 0.7589
Epoch 19/50
63/63 [==============================] - 0s 2ms/step - loss: 0.6389 - accuracy: 0.6836 - val_loss: 0.6769 - val_accuracy: 0.6786
Epoch 20/50
63/63 [==============================] - 0s 2ms/step - loss: 0.7437 - accuracy: 0.6948 - val_loss: 0.8652 - val_accuracy: 0.6250
Epoch 21/50
63/63 [==============================] - 0s 2ms/step - loss: 0.6828 - accuracy: 0.6741 - val_loss: 0.6399 - val_accuracy: 0.6875
Epoch 22/50
63/63 [==============================] - 0s 2ms/step - loss: 0.6228 - accuracy: 0.7202 - val_loss: 0.6255 - val_accuracy: 0.7143
Epoch 23/50
63/63 [==============================] - 0s 2ms/step - loss: 0.6451 - accuracy: 0.7138 - val_loss: 0.6170 - val_accuracy: 0.7321
Epoch 24/50
63/63 [==============================] - 0s 2ms/step - loss: 0.6931 - accuracy: 0.6884 - val_loss: 0.9736 - val_accuracy: 0.5625
Epoch 25/50
63/63 [==============================] - 0s 2ms/step - loss: 0.7165 - accuracy: 0.6804 - val_loss: 0.8022 - val_accuracy: 0.5982
Epoch 26/50
63/63 [==============================] - 0s 2ms/step - loss: 0.6296 - accuracy: 0.7218 - val_loss: 0.6108 - val_accuracy: 0.7589
Epoch 27/50
63/63 [==============================] - 0s 2ms/step - loss: 0.6003 - accuracy: 0.7281 - val_loss: 0.6273 - val_accuracy: 0.6964
Epoch 28/50
63/63 [==============================] - 0s 2ms/step - loss: 0.6128 - accuracy: 0.7361 - val_loss: 0.5741 - val_accuracy: 0.7143
Epoch 29/50
63/63 [==============================] - 0s 2ms/step - loss: 0.5718 - accuracy: 0.7409 - val_loss: 0.5542 - val_accuracy: 0.7500
Epoch 30/50
63/63 [==============================] - 0s 2ms/step - loss: 0.6038 - accuracy: 0.7472 - val_loss: 0.6641 - val_accuracy: 0.7143
Epoch 31/50
63/63 [==============================] - 0s 2ms/step - loss: 0.9322 - accuracy: 0.6312 - val_loss: 0.7455 - val_accuracy: 0.6250
Epoch 32/50
63/63 [==============================] - 0s 2ms/step - loss: 0.5835 - accuracy: 0.7647 - val_loss: 0.5372 - val_accuracy: 0.7679
Epoch 33/50
63/63 [==============================] - 0s 2ms/step - loss: 0.5335 - accuracy: 0.7647 - val_loss: 0.6049 - val_accuracy: 0.7321
Epoch 34/50
63/63 [==============================] - 0s 2ms/step - loss: 0.5938 - accuracy: 0.7329 - val_loss: 0.5833 - val_accuracy: 0.7857
Epoch 35/50
63/63 [==============================] - 0s 2ms/step - loss: 0.5931 - accuracy: 0.7361 - val_loss: 0.7136 - val_accuracy: 0.5982
Epoch 36/50
63/63 [==============================] - 0s 2ms/step - loss: 0.6114 - accuracy: 0.7393 - val_loss: 0.7923 - val_accuracy: 0.7143
Epoch 37/50
63/63 [==============================] - 0s 2ms/step - loss: 0.6299 - accuracy: 0.7456 - val_loss: 0.7031 - val_accuracy: 0.6250
Epoch 38/50
63/63 [==============================] - 0s 2ms/step - loss: 0.6340 - accuracy: 0.7202 - val_loss: 0.9424 - val_accuracy: 0.5714
Epoch 39/50
63/63 [==============================] - 0s 2ms/step - loss: 0.6644 - accuracy: 0.7043 - val_loss: 0.5474 - val_accuracy: 0.7768
Epoch 40/50
63/63 [==============================] - 0s 2ms/step - loss: 0.6125 - accuracy: 0.7472 - val_loss: 0.5849 - val_accuracy: 0.6875
Epoch 41/50
63/63 [==============================] - 0s 2ms/step - loss: 0.7174 - accuracy: 0.6979 - val_loss: 0.6808 - val_accuracy: 0.6429
Epoch 42/50
63/63 [==============================] - 0s 2ms/step - loss: 0.5745 - accuracy: 0.7234 - val_loss: 0.6556 - val_accuracy: 0.6607
Epoch 43/50
63/63 [==============================] - 0s 2ms/step - loss: 0.5407 - accuracy: 0.7663 - val_loss: 0.5297 - val_accuracy: 0.7679
Epoch 44/50
63/63 [==============================] - 0s 2ms/step - loss: 0.6271 - accuracy: 0.7297 - val_loss: 0.5195 - val_accuracy: 0.7857
Epoch 45/50
63/63 [==============================] - 0s 2ms/step - loss: 0.5284 - accuracy: 0.7568 - val_loss: 0.5185 - val_accuracy: 0.7143
Epoch 46/50
63/63 [==============================] - 0s 2ms/step - loss: 0.4938 - accuracy: 0.7790 - val_loss: 0.7390 - val_accuracy: 0.5714
Epoch 47/50
63/63 [==============================] - 0s 2ms/step - loss: 0.5497 - accuracy: 0.7663 - val_loss: 0.5103 - val_accuracy: 0.7768
Epoch 48/50
63/63 [==============================] - 0s 2ms/step - loss: 0.4905 - accuracy: 0.7790 - val_loss: 0.5439 - val_accuracy: 0.7500
Epoch 49/50
63/63 [==============================] - 0s 2ms/step - loss: 0.4659 - accuracy: 0.7981 - val_loss: 0.5015 - val_accuracy: 0.7321
Epoch 50/50
63/63 [==============================] - 0s 2ms/step - loss: 0.4882 - accuracy: 0.7886 - val_loss: 0.4462 - val_accuracy: 0.8393
10/10 [==============================] - 0s 1ms/step - loss: 0.4030 - accuracy: 0.8208
accuracy: 82.0755%
In [22]:
train_eval_model(X_train, y_train, X_test, y_test, 70, 20, 50, 5)

Epoch 1/50
126/126 [==============================] - 1s 3ms/step - loss: 2.2969 - accuracy: 0.4658 - val_loss: 1.0942 - val_accuracy: 0.3482
Epoch 2/50
126/126 [==============================] - 0s 2ms/step - loss: 1.2946 - accuracy: 0.4897 - val_loss: 1.3835 - val_accuracy: 0.4554
Epoch 3/50
126/126 [==============================] - 0s 2ms/step - loss: 1.0316 - accuracy: 0.5517 - val_loss: 1.1133 - val_accuracy: 0.5446
Epoch 4/50
126/126 [==============================] - 0s 2ms/step - loss: 1.0053 - accuracy: 0.5405 - val_loss: 1.5301 - val_accuracy: 0.3661
Epoch 5/50
126/126 [==============================] - 0s 2ms/step - loss: 1.1976 - accuracy: 0.4992 - val_loss: 1.3718 - val_accuracy: 0.5268
Epoch 6/50
126/126 [==============================] - 0s 2ms/step - loss: 0.9976 - accuracy: 0.6041 - val_loss: 0.8587 - val_accuracy: 0.6071
Epoch 7/50
126/126 [==============================] - 0s 2ms/step - loss: 0.9875 - accuracy: 0.5612 - val_loss: 0.8082 - val_accuracy: 0.5357
Epoch 8/50
126/126 [==============================] - 0s 2ms/step - loss: 0.9743 - accuracy: 0.6073 - val_loss: 1.2549 - val_accuracy: 0.6161
Epoch 9/50
126/126 [==============================] - 0s 2ms/step - loss: 1.0828 - accuracy: 0.5739 - val_loss: 1.2670 - val_accuracy: 0.6071
Epoch 10/50
126/126 [==============================] - 0s 2ms/step - loss: 0.8257 - accuracy: 0.6407 - val_loss: 0.7619 - val_accuracy: 0.6429
Epoch 11/50
126/126 [==============================] - 0s 2ms/step - loss: 0.8343 - accuracy: 0.6518 - val_loss: 0.9719 - val_accuracy: 0.4554
Epoch 12/50
126/126 [==============================] - 0s 2ms/step - loss: 0.7716 - accuracy: 0.6566 - val_loss: 0.8714 - val_accuracy: 0.5714
Epoch 13/50
126/126 [==============================] - 0s 2ms/step - loss: 0.8616 - accuracy: 0.6439 - val_loss: 0.9842 - val_accuracy: 0.5089
Epoch 14/50
126/126 [==============================] - 0s 2ms/step - loss: 0.8348 - accuracy: 0.6566 - val_loss: 0.9364 - val_accuracy: 0.5804
Epoch 15/50
126/126 [==============================] - 0s 2ms/step - loss: 0.9033 - accuracy: 0.6359 - val_loss: 1.0993 - val_accuracy: 0.6071
Epoch 16/50
126/126 [==============================] - 0s 2ms/step - loss: 0.8557 - accuracy: 0.6757 - val_loss: 0.7676 - val_accuracy: 0.7589
Epoch 17/50
126/126 [==============================] - 0s 1ms/step - loss: 0.9861 - accuracy: 0.6343 - val_loss: 1.1690 - val_accuracy: 0.5536
Epoch 18/50
126/126 [==============================] - 0s 2ms/step - loss: 0.7353 - accuracy: 0.6661 - val_loss: 0.8854 - val_accuracy: 0.4821
Epoch 19/50
126/126 [==============================] - 0s 2ms/step - loss: 0.7780 - accuracy: 0.6693 - val_loss: 0.7520 - val_accuracy: 0.6786
Epoch 20/50
126/126 [==============================] - 0s 2ms/step - loss: 0.7692 - accuracy: 0.6550 - val_loss: 1.5458 - val_accuracy: 0.5000
Epoch 21/50
126/126 [==============================] - 0s 2ms/step - loss: 0.7443 - accuracy: 0.6836 - val_loss: 0.9122 - val_accuracy: 0.5625
Epoch 22/50
126/126 [==============================] - 0s 2ms/step - loss: 0.7676 - accuracy: 0.6693 - val_loss: 0.7744 - val_accuracy: 0.6786
Epoch 23/50
126/126 [==============================] - 0s 2ms/step - loss: 0.9857 - accuracy: 0.6280 - val_loss: 0.6674 - val_accuracy: 0.7589
Epoch 24/50
126/126 [==============================] - 0s 2ms/step - loss: 0.6180 - accuracy: 0.7154 - val_loss: 0.6384 - val_accuracy: 0.6875
Epoch 25/50
126/126 [==============================] - 0s 2ms/step - loss: 0.6265 - accuracy: 0.7345 - val_loss: 1.1862 - val_accuracy: 0.5536
Epoch 26/50
126/126 [==============================] - 0s 2ms/step - loss: 0.7139 - accuracy: 0.7011 - val_loss: 0.7768 - val_accuracy: 0.6339
Epoch 27/50
126/126 [==============================] - 0s 2ms/step - loss: 0.6417 - accuracy: 0.7186 - val_loss: 0.8193 - val_accuracy: 0.6964
Epoch 28/50
126/126 [==============================] - 0s 2ms/step - loss: 0.6751 - accuracy: 0.7186 - val_loss: 1.0391 - val_accuracy: 0.6071
Epoch 29/50
126/126 [==============================] - 0s 2ms/step - loss: 0.6137 - accuracy: 0.7122 - val_loss: 0.5784 - val_accuracy: 0.7589
Epoch 30/50
126/126 [==============================] - 0s 2ms/step - loss: 0.5978 - accuracy: 0.7536 - val_loss: 0.6572 - val_accuracy: 0.6429
Epoch 31/50
126/126 [==============================] - 0s 2ms/step - loss: 0.7032 - accuracy: 0.6852 - val_loss: 0.7599 - val_accuracy: 0.6518
Epoch 32/50
126/126 [==============================] - 0s 2ms/step - loss: 0.6265 - accuracy: 0.7409 - val_loss: 0.6630 - val_accuracy: 0.6429
Epoch 33/50
126/126 [==============================] - 0s 2ms/step - loss: 0.5831 - accuracy: 0.7234 - val_loss: 0.7104 - val_accuracy: 0.7500
Epoch 34/50
126/126 [==============================] - 0s 2ms/step - loss: 0.5484 - accuracy: 0.7520 - val_loss: 0.9040 - val_accuracy: 0.6875
Epoch 35/50
126/126 [==============================] - 0s 2ms/step - loss: 0.7247 - accuracy: 0.7027 - val_loss: 0.8677 - val_accuracy: 0.6607
Epoch 36/50
126/126 [==============================] - 0s 2ms/step - loss: 0.6097 - accuracy: 0.7568 - val_loss: 0.5553 - val_accuracy: 0.7946
Epoch 37/50
126/126 [==============================] - 0s 2ms/step - loss: 0.6167 - accuracy: 0.7504 - val_loss: 0.6844 - val_accuracy: 0.6518
Epoch 38/50
126/126 [==============================] - 0s 2ms/step - loss: 0.5720 - accuracy: 0.7631 - val_loss: 0.9864 - val_accuracy: 0.6429
Epoch 39/50
126/126 [==============================] - 0s 2ms/step - loss: 0.5315 - accuracy: 0.7838 - val_loss: 0.7744 - val_accuracy: 0.7054
Epoch 40/50
126/126 [==============================] - 0s 2ms/step - loss: 0.5787 - accuracy: 0.7583 - val_loss: 1.2071 - val_accuracy: 0.6696
Epoch 41/50
126/126 [==============================] - 0s 2ms/step - loss: 0.5392 - accuracy: 0.7727 - val_loss: 0.8501 - val_accuracy: 0.6696
Epoch 42/50
126/126 [==============================] - 0s 2ms/step - loss: 0.6014 - accuracy: 0.7456 - val_loss: 0.8210 - val_accuracy: 0.6250
Epoch 43/50
126/126 [==============================] - 0s 2ms/step - loss: 0.5601 - accuracy: 0.7822 - val_loss: 0.8997 - val_accuracy: 0.6518
Epoch 44/50
126/126 [==============================] - 0s 2ms/step - loss: 0.6416 - accuracy: 0.7409 - val_loss: 0.5710 - val_accuracy: 0.8214
Epoch 45/50
126/126 [==============================] - 0s 2ms/step - loss: 0.5283 - accuracy: 0.7663 - val_loss: 0.5559 - val_accuracy: 0.7857
Epoch 46/50
126/126 [==============================] - 0s 2ms/step - loss: 0.4936 - accuracy: 0.8045 - val_loss: 0.5160 - val_accuracy: 0.7946
Epoch 47/50
126/126 [==============================] - 0s 2ms/step - loss: 0.4995 - accuracy: 0.7901 - val_loss: 0.4545 - val_accuracy: 0.7411
Epoch 48/50
126/126 [==============================] - 0s 2ms/step - loss: 0.5199 - accuracy: 0.7631 - val_loss: 0.5730 - val_accuracy: 0.7143
Epoch 49/50
126/126 [==============================] - 0s 2ms/step - loss: 0.4907 - accuracy: 0.7997 - val_loss: 0.4100 - val_accuracy: 0.7946
Epoch 50/50
126/126 [==============================] - 0s 2ms/step - loss: 0.5715 - accuracy: 0.7870 - val_loss: 0.5409 - val_accuracy: 0.7500
10/10 [==============================] - 0s 1ms/step - loss: 0.4496 - accuracy: 0.7956
accuracy: 79.5597%
In [23]:
train_eval_model(X_train, y_train, X_test, y_test, 50, 20, 50, 10)

Epoch 1/50
63/63 [==============================] - 1s 4ms/step - loss: 6.9141 - accuracy: 0.3545 - val_loss: 1.0747 - val_accuracy: 0.4732
Epoch 2/50
63/63 [==============================] - 0s 2ms/step - loss: 1.0428 - accuracy: 0.4626 - val_loss: 0.9854 - val_accuracy: 0.4911
Epoch 3/50
63/63 [==============================] - 0s 2ms/step - loss: 1.0224 - accuracy: 0.4754 - val_loss: 1.0577 - val_accuracy: 0.5089
Epoch 4/50
63/63 [==============================] - 0s 2ms/step - loss: 1.0067 - accuracy: 0.5151 - val_loss: 1.0302 - val_accuracy: 0.4732
Epoch 5/50
63/63 [==============================] - 0s 2ms/step - loss: 0.9409 - accuracy: 0.5421 - val_loss: 0.9339 - val_accuracy: 0.5536
Epoch 6/50
63/63 [==============================] - 0s 2ms/step - loss: 0.9411 - accuracy: 0.5596 - val_loss: 1.1926 - val_accuracy: 0.4643
Epoch 7/50
63/63 [==============================] - 0s 2ms/step - loss: 0.9366 - accuracy: 0.5644 - val_loss: 0.9623 - val_accuracy: 0.6339
Epoch 8/50
63/63 [==============================] - 0s 2ms/step - loss: 0.9401 - accuracy: 0.5119 - val_loss: 1.1667 - val_accuracy: 0.4643
Epoch 9/50
63/63 [==============================] - 0s 2ms/step - loss: 0.8929 - accuracy: 0.5692 - val_loss: 1.0977 - val_accuracy: 0.4821
Epoch 10/50
63/63 [==============================] - 0s 2ms/step - loss: 0.8649 - accuracy: 0.5851 - val_loss: 0.9227 - val_accuracy: 0.5446
Epoch 11/50
63/63 [==============================] - 0s 2ms/step - loss: 0.8892 - accuracy: 0.5803 - val_loss: 0.8851 - val_accuracy: 0.5625
Epoch 12/50
63/63 [==============================] - 0s 2ms/step - loss: 0.8291 - accuracy: 0.6025 - val_loss: 0.9172 - val_accuracy: 0.6607
Epoch 13/50
63/63 [==============================] - 0s 2ms/step - loss: 0.8467 - accuracy: 0.6169 - val_loss: 0.9739 - val_accuracy: 0.5179
Epoch 14/50
63/63 [==============================] - 0s 2ms/step - loss: 0.8958 - accuracy: 0.5819 - val_loss: 0.9321 - val_accuracy: 0.5268
Epoch 15/50
63/63 [==============================] - 0s 2ms/step - loss: 0.8132 - accuracy: 0.6486 - val_loss: 0.9210 - val_accuracy: 0.7054
Epoch 16/50
63/63 [==============================] - 0s 2ms/step - loss: 0.8182 - accuracy: 0.6248 - val_loss: 0.8460 - val_accuracy: 0.6607
Epoch 17/50
63/63 [==============================] - 0s 2ms/step - loss: 0.7993 - accuracy: 0.6312 - val_loss: 0.8292 - val_accuracy: 0.5982
Epoch 18/50
63/63 [==============================] - 0s 2ms/step - loss: 0.7896 - accuracy: 0.6693 - val_loss: 0.8175 - val_accuracy: 0.5714
Epoch 19/50
63/63 [==============================] - 0s 2ms/step - loss: 0.7814 - accuracy: 0.6645 - val_loss: 0.8240 - val_accuracy: 0.6250
Epoch 20/50
63/63 [==============================] - 0s 2ms/step - loss: 0.7454 - accuracy: 0.6773 - val_loss: 0.8625 - val_accuracy: 0.6696
Epoch 21/50
63/63 [==============================] - 0s 2ms/step - loss: 0.7623 - accuracy: 0.6645 - val_loss: 0.8049 - val_accuracy: 0.6161
Epoch 22/50
63/63 [==============================] - 0s 2ms/step - loss: 0.7258 - accuracy: 0.6836 - val_loss: 0.7980 - val_accuracy: 0.6161
Epoch 23/50
63/63 [==============================] - 0s 2ms/step - loss: 0.7693 - accuracy: 0.6582 - val_loss: 0.8935 - val_accuracy: 0.7054
Epoch 24/50
63/63 [==============================] - 0s 2ms/step - loss: 0.7948 - accuracy: 0.6709 - val_loss: 0.8584 - val_accuracy: 0.7500
Epoch 25/50
63/63 [==============================] - 0s 2ms/step - loss: 0.7397 - accuracy: 0.6725 - val_loss: 0.7877 - val_accuracy: 0.5893
Epoch 26/50
63/63 [==============================] - 0s 2ms/step - loss: 0.7131 - accuracy: 0.6868 - val_loss: 0.8571 - val_accuracy: 0.5714
Epoch 27/50
63/63 [==============================] - 0s 2ms/step - loss: 0.7725 - accuracy: 0.6677 - val_loss: 0.7583 - val_accuracy: 0.6250
Epoch 28/50
63/63 [==============================] - 0s 2ms/step - loss: 0.6972 - accuracy: 0.6948 - val_loss: 0.8489 - val_accuracy: 0.5804
Epoch 29/50
63/63 [==============================] - 0s 2ms/step - loss: 0.6845 - accuracy: 0.6979 - val_loss: 0.7653 - val_accuracy: 0.6964
Epoch 30/50
63/63 [==============================] - 0s 2ms/step - loss: 0.6801 - accuracy: 0.6804 - val_loss: 0.8307 - val_accuracy: 0.5714
Epoch 31/50
63/63 [==============================] - 0s 2ms/step - loss: 0.6581 - accuracy: 0.6979 - val_loss: 0.7604 - val_accuracy: 0.6786
Epoch 32/50
63/63 [==============================] - 0s 2ms/step - loss: 0.6988 - accuracy: 0.6773 - val_loss: 0.7699 - val_accuracy: 0.5893
Epoch 33/50
63/63 [==============================] - 0s 2ms/step - loss: 0.6617 - accuracy: 0.7011 - val_loss: 0.7974 - val_accuracy: 0.7232
Epoch 34/50
63/63 [==============================] - 0s 2ms/step - loss: 0.7195 - accuracy: 0.6789 - val_loss: 0.7458 - val_accuracy: 0.7589
Epoch 35/50
63/63 [==============================] - 0s 2ms/step - loss: 0.6325 - accuracy: 0.7377 - val_loss: 0.7642 - val_accuracy: 0.6518
Epoch 36/50
63/63 [==============================] - 0s 2ms/step - loss: 0.6250 - accuracy: 0.7281 - val_loss: 0.6723 - val_accuracy: 0.6607
Epoch 37/50
63/63 [==============================] - 0s 2ms/step - loss: 0.6287 - accuracy: 0.7234 - val_loss: 0.6813 - val_accuracy: 0.6607
Epoch 38/50
63/63 [==============================] - 0s 2ms/step - loss: 0.7276 - accuracy: 0.6693 - val_loss: 0.7567 - val_accuracy: 0.7500
Epoch 39/50
63/63 [==============================] - 0s 2ms/step - loss: 0.6268 - accuracy: 0.7250 - val_loss: 0.7085 - val_accuracy: 0.7143
Epoch 40/50
63/63 [==============================] - 0s 2ms/step - loss: 0.5967 - accuracy: 0.7345 - val_loss: 0.6490 - val_accuracy: 0.7500
Epoch 41/50
63/63 [==============================] - 0s 2ms/step - loss: 0.5830 - accuracy: 0.7615 - val_loss: 0.7144 - val_accuracy: 0.7143
Epoch 42/50
63/63 [==============================] - 0s 2ms/step - loss: 0.6200 - accuracy: 0.7393 - val_loss: 0.8224 - val_accuracy: 0.5982
Epoch 43/50
63/63 [==============================] - 0s 2ms/step - loss: 0.6009 - accuracy: 0.7409 - val_loss: 0.6548 - val_accuracy: 0.7143
Epoch 44/50
63/63 [==============================] - 0s 2ms/step - loss: 0.5935 - accuracy: 0.7409 - val_loss: 0.6259 - val_accuracy: 0.7679
Epoch 45/50
63/63 [==============================] - 0s 2ms/step - loss: 0.6194 - accuracy: 0.7202 - val_loss: 0.6364 - val_accuracy: 0.7500
Epoch 46/50
63/63 [==============================] - 0s 2ms/step - loss: 0.6487 - accuracy: 0.6836 - val_loss: 0.6992 - val_accuracy: 0.6339
Epoch 47/50
63/63 [==============================] - 0s 2ms/step - loss: 0.5871 - accuracy: 0.7440 - val_loss: 0.6494 - val_accuracy: 0.7500
Epoch 48/50
63/63 [==============================] - 0s 2ms/step - loss: 0.6030 - accuracy: 0.7361 - val_loss: 0.6092 - val_accuracy: 0.7500
Epoch 49/50
63/63 [==============================] - 0s 2ms/step - loss: 0.6058 - accuracy: 0.7266 - val_loss: 0.7162 - val_accuracy: 0.7232
Epoch 50/50
63/63 [==============================] - 0s 2ms/step - loss: 0.6546 - accuracy: 0.7075 - val_loss: 0.7792 - val_accuracy: 0.6964
10/10 [==============================] - 0s 1ms/step - loss: 0.6935 - accuracy: 0.6981
accuracy: 69.8113%
In [26]:
train_eval_model(X_train, y_train, X_test, y_test, 70, 30, 50, 10)

Epoch 1/50
63/63 [==============================] - 1s 4ms/step - loss: 1.2850 - accuracy: 0.4785 - val_loss: 1.0657 - val_accuracy: 0.5357
Epoch 2/50
63/63 [==============================] - 0s 2ms/step - loss: 1.0277 - accuracy: 0.5278 - val_loss: 0.9072 - val_accuracy: 0.5536
Epoch 3/50
63/63 [==============================] - 0s 2ms/step - loss: 1.1075 - accuracy: 0.5580 - val_loss: 1.1300 - val_accuracy: 0.5714
Epoch 4/50
63/63 [==============================] - 0s 2ms/step - loss: 0.9652 - accuracy: 0.5676 - val_loss: 0.8283 - val_accuracy: 0.6161
Epoch 5/50
63/63 [==============================] - 0s 2ms/step - loss: 0.9778 - accuracy: 0.6264 - val_loss: 0.7938 - val_accuracy: 0.6250
Epoch 6/50
63/63 [==============================] - 0s 2ms/step - loss: 0.7436 - accuracy: 0.6534 - val_loss: 0.7402 - val_accuracy: 0.6875
Epoch 7/50
63/63 [==============================] - 0s 2ms/step - loss: 0.9680 - accuracy: 0.5739 - val_loss: 2.6478 - val_accuracy: 0.5982
Epoch 8/50
63/63 [==============================] - 0s 2ms/step - loss: 1.0699 - accuracy: 0.6025 - val_loss: 1.0171 - val_accuracy: 0.5536
Epoch 9/50
63/63 [==============================] - 0s 2ms/step - loss: 0.7589 - accuracy: 0.6598 - val_loss: 0.7226 - val_accuracy: 0.6696
Epoch 10/50
63/63 [==============================] - 0s 2ms/step - loss: 0.7224 - accuracy: 0.6550 - val_loss: 0.6754 - val_accuracy: 0.7500
Epoch 11/50
63/63 [==============================] - 0s 2ms/step - loss: 0.7915 - accuracy: 0.6661 - val_loss: 0.6432 - val_accuracy: 0.7232
Epoch 12/50
63/63 [==============================] - 0s 2ms/step - loss: 0.7140 - accuracy: 0.6916 - val_loss: 0.7559 - val_accuracy: 0.6339
Epoch 13/50
63/63 [==============================] - 0s 2ms/step - loss: 0.7792 - accuracy: 0.6741 - val_loss: 0.7355 - val_accuracy: 0.7321
Epoch 14/50
63/63 [==============================] - 0s 2ms/step - loss: 0.6354 - accuracy: 0.7250 - val_loss: 0.6398 - val_accuracy: 0.6786
Epoch 15/50
63/63 [==============================] - 0s 2ms/step - loss: 0.6683 - accuracy: 0.7011 - val_loss: 0.8241 - val_accuracy: 0.5804
Epoch 16/50
63/63 [==============================] - 0s 2ms/step - loss: 0.9224 - accuracy: 0.6328 - val_loss: 0.6183 - val_accuracy: 0.7589
Epoch 17/50
63/63 [==============================] - 0s 2ms/step - loss: 0.7452 - accuracy: 0.6852 - val_loss: 0.8730 - val_accuracy: 0.5982
Epoch 18/50
63/63 [==============================] - 0s 2ms/step - loss: 0.6585 - accuracy: 0.6773 - val_loss: 0.5799 - val_accuracy: 0.6161
Epoch 19/50
63/63 [==============================] - 0s 2ms/step - loss: 0.6210 - accuracy: 0.7122 - val_loss: 0.5767 - val_accuracy: 0.7589
Epoch 20/50
63/63 [==============================] - 0s 2ms/step - loss: 0.6530 - accuracy: 0.7186 - val_loss: 0.5541 - val_accuracy: 0.7857
Epoch 21/50
63/63 [==============================] - 0s 2ms/step - loss: 0.6533 - accuracy: 0.7043 - val_loss: 0.5661 - val_accuracy: 0.7143
Epoch 22/50
63/63 [==============================] - 0s 2ms/step - loss: 0.6405 - accuracy: 0.7107 - val_loss: 0.6083 - val_accuracy: 0.6161
Epoch 23/50
63/63 [==============================] - 0s 2ms/step - loss: 0.6114 - accuracy: 0.7297 - val_loss: 0.5235 - val_accuracy: 0.7768
Epoch 24/50
63/63 [==============================] - 0s 2ms/step - loss: 0.6343 - accuracy: 0.7027 - val_loss: 0.5130 - val_accuracy: 0.7768
Epoch 25/50
63/63 [==============================] - 0s 2ms/step - loss: 0.6415 - accuracy: 0.6852 - val_loss: 0.7315 - val_accuracy: 0.6786
Epoch 26/50
63/63 [==============================] - 0s 2ms/step - loss: 0.5825 - accuracy: 0.7281 - val_loss: 0.6207 - val_accuracy: 0.6607
Epoch 27/50
63/63 [==============================] - 0s 2ms/step - loss: 0.6347 - accuracy: 0.7234 - val_loss: 0.6045 - val_accuracy: 0.6696
Epoch 28/50
63/63 [==============================] - 0s 2ms/step - loss: 0.7236 - accuracy: 0.7122 - val_loss: 0.6759 - val_accuracy: 0.7143
Epoch 29/50
63/63 [==============================] - 0s 2ms/step - loss: 0.6271 - accuracy: 0.7122 - val_loss: 0.8235 - val_accuracy: 0.5804
Epoch 30/50
63/63 [==============================] - 0s 2ms/step - loss: 0.5849 - accuracy: 0.7250 - val_loss: 0.5284 - val_accuracy: 0.7768
Epoch 31/50
63/63 [==============================] - 0s 2ms/step - loss: 0.5001 - accuracy: 0.7758 - val_loss: 0.5528 - val_accuracy: 0.7500
Epoch 32/50
63/63 [==============================] - 0s 2ms/step - loss: 0.5198 - accuracy: 0.7520 - val_loss: 0.7127 - val_accuracy: 0.6875
Epoch 33/50
63/63 [==============================] - 0s 2ms/step - loss: 0.5207 - accuracy: 0.7615 - val_loss: 0.4854 - val_accuracy: 0.7679
Epoch 34/50
63/63 [==============================] - 0s 2ms/step - loss: 0.5396 - accuracy: 0.7552 - val_loss: 0.4613 - val_accuracy: 0.7857
Epoch 35/50
63/63 [==============================] - 0s 2ms/step - loss: 0.7100 - accuracy: 0.7250 - val_loss: 0.6244 - val_accuracy: 0.7232
Epoch 36/50
63/63 [==============================] - 0s 2ms/step - loss: 0.4870 - accuracy: 0.7806 - val_loss: 0.4958 - val_accuracy: 0.6786
Epoch 37/50
63/63 [==============================] - 0s 2ms/step - loss: 0.4635 - accuracy: 0.7806 - val_loss: 0.5129 - val_accuracy: 0.6964
Epoch 38/50
63/63 [==============================] - 0s 2ms/step - loss: 0.4842 - accuracy: 0.7933 - val_loss: 0.4889 - val_accuracy: 0.7500
Epoch 39/50
63/63 [==============================] - 0s 2ms/step - loss: 0.5184 - accuracy: 0.7679 - val_loss: 0.5545 - val_accuracy: 0.6964
Epoch 40/50
63/63 [==============================] - 0s 2ms/step - loss: 0.4808 - accuracy: 0.7727 - val_loss: 0.4919 - val_accuracy: 0.7500
Epoch 41/50
63/63 [==============================] - 0s 2ms/step - loss: 0.5093 - accuracy: 0.7695 - val_loss: 0.6251 - val_accuracy: 0.7054
Epoch 42/50
63/63 [==============================] - 0s 2ms/step - loss: 0.5745 - accuracy: 0.7504 - val_loss: 0.9578 - val_accuracy: 0.5982
Epoch 43/50
63/63 [==============================] - 0s 2ms/step - loss: 0.6014 - accuracy: 0.7599 - val_loss: 0.8229 - val_accuracy: 0.7232
Epoch 44/50
63/63 [==============================] - 0s 2ms/step - loss: 0.4658 - accuracy: 0.8013 - val_loss: 0.6169 - val_accuracy: 0.6875
Epoch 45/50
63/63 [==============================] - 0s 2ms/step - loss: 0.5121 - accuracy: 0.7727 - val_loss: 0.4469 - val_accuracy: 0.7768
Epoch 46/50
63/63 [==============================] - 0s 2ms/step - loss: 0.4361 - accuracy: 0.7933 - val_loss: 0.4424 - val_accuracy: 0.8304
Epoch 47/50
63/63 [==============================] - 0s 2ms/step - loss: 0.4278 - accuracy: 0.8108 - val_loss: 0.5071 - val_accuracy: 0.7232
Epoch 48/50
63/63 [==============================] - 0s 2ms/step - loss: 0.6029 - accuracy: 0.7568 - val_loss: 0.6997 - val_accuracy: 0.7143
Epoch 49/50
63/63 [==============================] - 0s 2ms/step - loss: 0.4248 - accuracy: 0.8060 - val_loss: 0.5877 - val_accuracy: 0.7768
Epoch 50/50
63/63 [==============================] - 0s 2ms/step - loss: 0.4842 - accuracy: 0.7774 - val_loss: 0.5205 - val_accuracy: 0.7768
10/10 [==============================] - 0s 1ms/step - loss: 0.4369 - accuracy: 0.8553
accuracy: 85.5346%
In [28]:
train_eval_model(X_train, y_train, X_test, y_test, 60, 30, 50, 15)

Epoch 1/50
42/42 [==============================] - 1s 6ms/step - loss: 7.8885 - accuracy: 0.3323 - val_loss: 2.1381 - val_accuracy: 0.4018
Epoch 2/50
42/42 [==============================] - 0s 2ms/step - loss: 1.2099 - accuracy: 0.4436 - val_loss: 1.0459 - val_accuracy: 0.4018
Epoch 3/50
42/42 [==============================] - 0s 2ms/step - loss: 0.9792 - accuracy: 0.5278 - val_loss: 1.0544 - val_accuracy: 0.4018
Epoch 4/50
42/42 [==============================] - 0s 2ms/step - loss: 0.9715 - accuracy: 0.5246 - val_loss: 0.9267 - val_accuracy: 0.5089
Epoch 5/50
42/42 [==============================] - 0s 2ms/step - loss: 0.9324 - accuracy: 0.5469 - val_loss: 0.9090 - val_accuracy: 0.6518
Epoch 6/50
42/42 [==============================] - 0s 2ms/step - loss: 0.8760 - accuracy: 0.5819 - val_loss: 0.9547 - val_accuracy: 0.4286
Epoch 7/50
42/42 [==============================] - 0s 2ms/step - loss: 0.9140 - accuracy: 0.5548 - val_loss: 0.9405 - val_accuracy: 0.6250
Epoch 8/50
42/42 [==============================] - 0s 2ms/step - loss: 0.9416 - accuracy: 0.5580 - val_loss: 0.8651 - val_accuracy: 0.6696
Epoch 9/50
42/42 [==============================] - 0s 2ms/step - loss: 1.0429 - accuracy: 0.5199 - val_loss: 1.6204 - val_accuracy: 0.4286
Epoch 10/50
42/42 [==============================] - 0s 2ms/step - loss: 1.0356 - accuracy: 0.5453 - val_loss: 0.9599 - val_accuracy: 0.5089
Epoch 11/50
42/42 [==============================] - 0s 2ms/step - loss: 0.9659 - accuracy: 0.5660 - val_loss: 0.8787 - val_accuracy: 0.5625
Epoch 12/50
42/42 [==============================] - 0s 2ms/step - loss: 0.7957 - accuracy: 0.6502 - val_loss: 0.8323 - val_accuracy: 0.7411
Epoch 13/50
42/42 [==============================] - 0s 2ms/step - loss: 0.8448 - accuracy: 0.6121 - val_loss: 0.8539 - val_accuracy: 0.5804
Epoch 14/50
42/42 [==============================] - 0s 2ms/step - loss: 0.7919 - accuracy: 0.6423 - val_loss: 0.8117 - val_accuracy: 0.5625
Epoch 15/50
42/42 [==============================] - 0s 2ms/step - loss: 0.8425 - accuracy: 0.6089 - val_loss: 1.0041 - val_accuracy: 0.5804
Epoch 16/50
42/42 [==============================] - 0s 2ms/step - loss: 0.8978 - accuracy: 0.5628 - val_loss: 0.8510 - val_accuracy: 0.6786
Epoch 17/50
42/42 [==============================] - 0s 2ms/step - loss: 0.7469 - accuracy: 0.6614 - val_loss: 0.7746 - val_accuracy: 0.7589
Epoch 18/50
42/42 [==============================] - 0s 2ms/step - loss: 0.8309 - accuracy: 0.6137 - val_loss: 0.7890 - val_accuracy: 0.5982
Epoch 19/50
42/42 [==============================] - 0s 2ms/step - loss: 0.7421 - accuracy: 0.6852 - val_loss: 0.7620 - val_accuracy: 0.7768
Epoch 20/50
42/42 [==============================] - 0s 2ms/step - loss: 0.8277 - accuracy: 0.6343 - val_loss: 1.1193 - val_accuracy: 0.5804
Epoch 21/50
42/42 [==============================] - 0s 2ms/step - loss: 0.6891 - accuracy: 0.7091 - val_loss: 0.7698 - val_accuracy: 0.6250
Epoch 22/50
42/42 [==============================] - 0s 2ms/step - loss: 0.7214 - accuracy: 0.6804 - val_loss: 0.9924 - val_accuracy: 0.5357
Epoch 23/50
42/42 [==============================] - 0s 2ms/step - loss: 0.7227 - accuracy: 0.6868 - val_loss: 0.7633 - val_accuracy: 0.6071
Epoch 24/50
42/42 [==============================] - 0s 2ms/step - loss: 0.6642 - accuracy: 0.7107 - val_loss: 0.7352 - val_accuracy: 0.6696
Epoch 25/50
42/42 [==============================] - 0s 2ms/step - loss: 0.7192 - accuracy: 0.6582 - val_loss: 0.7395 - val_accuracy: 0.6339
Epoch 26/50
42/42 [==============================] - 0s 2ms/step - loss: 0.6899 - accuracy: 0.6979 - val_loss: 0.8213 - val_accuracy: 0.6339
Epoch 27/50
42/42 [==============================] - 0s 2ms/step - loss: 0.6750 - accuracy: 0.6932 - val_loss: 0.6917 - val_accuracy: 0.6161
Epoch 28/50
42/42 [==============================] - 0s 2ms/step - loss: 0.6120 - accuracy: 0.7456 - val_loss: 0.6846 - val_accuracy: 0.8036
Epoch 29/50
42/42 [==============================] - 0s 2ms/step - loss: 0.5950 - accuracy: 0.7917 - val_loss: 0.6754 - val_accuracy: 0.7768
Epoch 30/50
42/42 [==============================] - 0s 2ms/step - loss: 0.6948 - accuracy: 0.6916 - val_loss: 0.9082 - val_accuracy: 0.6429
Epoch 31/50
42/42 [==============================] - 0s 2ms/step - loss: 0.7332 - accuracy: 0.6709 - val_loss: 0.7486 - val_accuracy: 0.6339
Epoch 32/50
42/42 [==============================] - 0s 3ms/step - loss: 0.7167 - accuracy: 0.6789 - val_loss: 1.0037 - val_accuracy: 0.5179
Epoch 33/50
42/42 [==============================] - 0s 2ms/step - loss: 0.5918 - accuracy: 0.7631 - val_loss: 0.6520 - val_accuracy: 0.6429
Epoch 34/50
42/42 [==============================] - 0s 2ms/step - loss: 0.5657 - accuracy: 0.7663 - val_loss: 0.6412 - val_accuracy: 0.7500
Epoch 35/50
42/42 [==============================] - 0s 2ms/step - loss: 0.6055 - accuracy: 0.7313 - val_loss: 0.6909 - val_accuracy: 0.6875
Epoch 36/50
42/42 [==============================] - 0s 2ms/step - loss: 0.7574 - accuracy: 0.6820 - val_loss: 0.8222 - val_accuracy: 0.7054
Epoch 37/50
42/42 [==============================] - 0s 2ms/step - loss: 0.6898 - accuracy: 0.7138 - val_loss: 0.7027 - val_accuracy: 0.6071
Epoch 38/50
42/42 [==============================] - 0s 2ms/step - loss: 0.5568 - accuracy: 0.7615 - val_loss: 0.6485 - val_accuracy: 0.6786
Epoch 39/50
42/42 [==============================] - 0s 2ms/step - loss: 0.6895 - accuracy: 0.7250 - val_loss: 0.6597 - val_accuracy: 0.6518
Epoch 40/50
42/42 [==============================] - 0s 2ms/step - loss: 0.5271 - accuracy: 0.7949 - val_loss: 0.6562 - val_accuracy: 0.6518
Epoch 41/50
42/42 [==============================] - 0s 2ms/step - loss: 0.6379 - accuracy: 0.7297 - val_loss: 0.5836 - val_accuracy: 0.7857
Epoch 42/50
42/42 [==============================] - 0s 2ms/step - loss: 0.5768 - accuracy: 0.7536 - val_loss: 0.5847 - val_accuracy: 0.8036
Epoch 43/50
42/42 [==============================] - 0s 2ms/step - loss: 0.5439 - accuracy: 0.7568 - val_loss: 0.5535 - val_accuracy: 0.7768
Epoch 44/50
42/42 [==============================] - 0s 2ms/step - loss: 0.5176 - accuracy: 0.7806 - val_loss: 0.8357 - val_accuracy: 0.7321
Epoch 45/50
42/42 [==============================] - 0s 2ms/step - loss: 0.5289 - accuracy: 0.7599 - val_loss: 0.5671 - val_accuracy: 0.7768
Epoch 46/50
42/42 [==============================] - 0s 3ms/step - loss: 0.5255 - accuracy: 0.7774 - val_loss: 0.6534 - val_accuracy: 0.7054
Epoch 47/50
42/42 [==============================] - 0s 2ms/step - loss: 0.5702 - accuracy: 0.7679 - val_loss: 0.5472 - val_accuracy: 0.8125
Epoch 48/50
42/42 [==============================] - 0s 2ms/step - loss: 0.5781 - accuracy: 0.7409 - val_loss: 0.7194 - val_accuracy: 0.5804
Epoch 49/50
42/42 [==============================] - 0s 2ms/step - loss: 0.5438 - accuracy: 0.7663 - val_loss: 0.6366 - val_accuracy: 0.6786
Epoch 50/50
42/42 [==============================] - 0s 2ms/step - loss: 0.5338 - accuracy: 0.7806 - val_loss: 0.6490 - val_accuracy: 0.7946
10/10 [==============================] - 0s 1ms/step - loss: 0.5310 - accuracy: 0.8145
accuracy: 81.4465%
In [31]:
train_eval_model(X_train, y_train, X_test, y_test, 55, 30, 50, 15)

Epoch 1/50
42/42 [==============================] - 1s 5ms/step - loss: 17.8443 - accuracy: 0.3657 - val_loss: 1.4768 - val_accuracy: 0.4018
Epoch 2/50
42/42 [==============================] - 0s 2ms/step - loss: 1.2789 - accuracy: 0.4070 - val_loss: 1.1073 - val_accuracy: 0.4018
Epoch 3/50
42/42 [==============================] - 0s 2ms/step - loss: 1.0814 - accuracy: 0.4738 - val_loss: 1.2951 - val_accuracy: 0.4018
Epoch 4/50
42/42 [==============================] - 0s 2ms/step - loss: 1.1580 - accuracy: 0.4674 - val_loss: 1.5469 - val_accuracy: 0.2946
Epoch 5/50
42/42 [==============================] - 0s 2ms/step - loss: 1.2102 - accuracy: 0.4722 - val_loss: 1.4464 - val_accuracy: 0.3571
Epoch 6/50
42/42 [==============================] - 0s 2ms/step - loss: 1.0155 - accuracy: 0.5246 - val_loss: 0.9911 - val_accuracy: 0.6071
Epoch 7/50
42/42 [==============================] - 0s 2ms/step - loss: 0.8978 - accuracy: 0.5771 - val_loss: 0.9413 - val_accuracy: 0.5982
Epoch 8/50
42/42 [==============================] - 0s 2ms/step - loss: 0.8977 - accuracy: 0.6089 - val_loss: 0.9143 - val_accuracy: 0.4911
Epoch 9/50
42/42 [==============================] - 0s 2ms/step - loss: 0.8544 - accuracy: 0.5676 - val_loss: 0.9727 - val_accuracy: 0.5089
Epoch 10/50
42/42 [==============================] - 0s 2ms/step - loss: 0.8415 - accuracy: 0.6010 - val_loss: 0.9044 - val_accuracy: 0.6429
Epoch 11/50
42/42 [==============================] - 0s 2ms/step - loss: 0.9571 - accuracy: 0.5739 - val_loss: 1.0746 - val_accuracy: 0.5446
Epoch 12/50
42/42 [==============================] - 0s 2ms/step - loss: 0.8441 - accuracy: 0.6121 - val_loss: 0.8581 - val_accuracy: 0.5089
Epoch 13/50
42/42 [==============================] - 0s 2ms/step - loss: 0.9127 - accuracy: 0.5644 - val_loss: 0.9773 - val_accuracy: 0.4821
Epoch 14/50
42/42 [==============================] - 0s 2ms/step - loss: 0.8313 - accuracy: 0.6232 - val_loss: 1.0047 - val_accuracy: 0.4464
Epoch 15/50
42/42 [==============================] - 0s 2ms/step - loss: 0.8349 - accuracy: 0.6089 - val_loss: 0.7974 - val_accuracy: 0.7232
Epoch 16/50
42/42 [==============================] - 0s 2ms/step - loss: 0.8043 - accuracy: 0.6407 - val_loss: 0.9041 - val_accuracy: 0.7054
Epoch 17/50
42/42 [==============================] - 0s 2ms/step - loss: 0.7919 - accuracy: 0.6391 - val_loss: 0.7596 - val_accuracy: 0.6429
Epoch 18/50
42/42 [==============================] - 0s 2ms/step - loss: 0.8036 - accuracy: 0.6280 - val_loss: 0.7814 - val_accuracy: 0.6607
Epoch 19/50
42/42 [==============================] - 0s 2ms/step - loss: 0.7364 - accuracy: 0.6455 - val_loss: 0.7235 - val_accuracy: 0.6518
Epoch 20/50
42/42 [==============================] - 0s 2ms/step - loss: 0.7479 - accuracy: 0.6677 - val_loss: 0.7065 - val_accuracy: 0.6250
Epoch 21/50
42/42 [==============================] - 0s 2ms/step - loss: 0.6815 - accuracy: 0.7043 - val_loss: 0.7105 - val_accuracy: 0.7232
Epoch 22/50
42/42 [==============================] - 0s 2ms/step - loss: 0.7053 - accuracy: 0.6979 - val_loss: 0.7220 - val_accuracy: 0.7500
Epoch 23/50
42/42 [==============================] - 0s 3ms/step - loss: 0.7192 - accuracy: 0.6725 - val_loss: 0.7412 - val_accuracy: 0.5625
Epoch 24/50
42/42 [==============================] - 0s 2ms/step - loss: 0.6894 - accuracy: 0.6900 - val_loss: 0.8086 - val_accuracy: 0.6071
Epoch 25/50
42/42 [==============================] - 0s 2ms/step - loss: 0.7546 - accuracy: 0.6423 - val_loss: 0.6987 - val_accuracy: 0.7589
Epoch 26/50
42/42 [==============================] - 0s 2ms/step - loss: 0.6755 - accuracy: 0.7043 - val_loss: 0.7751 - val_accuracy: 0.6607
Epoch 27/50
42/42 [==============================] - 0s 2ms/step - loss: 0.6426 - accuracy: 0.7011 - val_loss: 0.8959 - val_accuracy: 0.5804
Epoch 28/50
42/42 [==============================] - 0s 2ms/step - loss: 0.6607 - accuracy: 0.7107 - val_loss: 0.7024 - val_accuracy: 0.7054
Epoch 29/50
42/42 [==============================] - 0s 2ms/step - loss: 0.6383 - accuracy: 0.7154 - val_loss: 0.6420 - val_accuracy: 0.7054
Epoch 30/50
42/42 [==============================] - 0s 2ms/step - loss: 0.6402 - accuracy: 0.7027 - val_loss: 0.7040 - val_accuracy: 0.7232
Epoch 31/50
42/42 [==============================] - 0s 2ms/step - loss: 0.6582 - accuracy: 0.6916 - val_loss: 0.6676 - val_accuracy: 0.6607
Epoch 32/50
42/42 [==============================] - 0s 2ms/step - loss: 0.6301 - accuracy: 0.6963 - val_loss: 0.6399 - val_accuracy: 0.7143
Epoch 33/50
42/42 [==============================] - 0s 2ms/step - loss: 0.6320 - accuracy: 0.7154 - val_loss: 0.6226 - val_accuracy: 0.7143
Epoch 34/50
42/42 [==============================] - 0s 2ms/step - loss: 0.6810 - accuracy: 0.7027 - val_loss: 0.7565 - val_accuracy: 0.6161
Epoch 35/50
42/42 [==============================] - 0s 2ms/step - loss: 0.6625 - accuracy: 0.7122 - val_loss: 0.6503 - val_accuracy: 0.6429
Epoch 36/50
42/42 [==============================] - 0s 2ms/step - loss: 0.6237 - accuracy: 0.7091 - val_loss: 0.6127 - val_accuracy: 0.7321
Epoch 37/50
42/42 [==============================] - 0s 2ms/step - loss: 0.6962 - accuracy: 0.6582 - val_loss: 0.7192 - val_accuracy: 0.6696
Epoch 38/50
42/42 [==============================] - 0s 2ms/step - loss: 0.6046 - accuracy: 0.7218 - val_loss: 0.7005 - val_accuracy: 0.7143
Epoch 39/50
42/42 [==============================] - 0s 2ms/step - loss: 0.6431 - accuracy: 0.6916 - val_loss: 0.6302 - val_accuracy: 0.7500
Epoch 40/50
42/42 [==============================] - 0s 2ms/step - loss: 0.6209 - accuracy: 0.7297 - val_loss: 0.6239 - val_accuracy: 0.6875
Epoch 41/50
42/42 [==============================] - 0s 2ms/step - loss: 0.5763 - accuracy: 0.7218 - val_loss: 0.6262 - val_accuracy: 0.7321
Epoch 42/50
42/42 [==============================] - 0s 2ms/step - loss: 0.5786 - accuracy: 0.7409 - val_loss: 0.6459 - val_accuracy: 0.6964
Epoch 43/50
42/42 [==============================] - 0s 2ms/step - loss: 0.6030 - accuracy: 0.7281 - val_loss: 0.6941 - val_accuracy: 0.7500
Epoch 44/50
42/42 [==============================] - 0s 2ms/step - loss: 0.5604 - accuracy: 0.7472 - val_loss: 0.6279 - val_accuracy: 0.7500
Epoch 45/50
42/42 [==============================] - 0s 2ms/step - loss: 0.6183 - accuracy: 0.7234 - val_loss: 0.7836 - val_accuracy: 0.6518
Epoch 46/50
42/42 [==============================] - 0s 2ms/step - loss: 0.6234 - accuracy: 0.7186 - val_loss: 0.6689 - val_accuracy: 0.6786
Epoch 47/50
42/42 [==============================] - 0s 2ms/step - loss: 0.5716 - accuracy: 0.7440 - val_loss: 0.6753 - val_accuracy: 0.7232
Epoch 48/50
42/42 [==============================] - 0s 2ms/step - loss: 0.5511 - accuracy: 0.7711 - val_loss: 0.5650 - val_accuracy: 0.7500
Epoch 49/50
42/42 [==============================] - 0s 2ms/step - loss: 0.5582 - accuracy: 0.7615 - val_loss: 0.8085 - val_accuracy: 0.5982
Epoch 50/50
42/42 [==============================] - 0s 2ms/step - loss: 0.6742 - accuracy: 0.6963 - val_loss: 0.5591 - val_accuracy: 0.7857
10/10 [==============================] - 0s 1ms/step - loss: 0.4754 - accuracy: 0.8208
accuracy: 82.0755%
In [32]:
train_eval_model(X_train, y_train, X_test, y_test, 55, 30, 50, 15)

Epoch 1/150
42/42 [==============================] - 1s 6ms/step - loss: 2.6267 - accuracy: 0.3895 - val_loss: 1.2122 - val_accuracy: 0.4643
Epoch 2/150
42/42 [==============================] - 0s 2ms/step - loss: 0.9439 - accuracy: 0.5533 - val_loss: 1.1734 - val_accuracy: 0.5804
Epoch 3/150
42/42 [==============================] - 0s 2ms/step - loss: 0.9321 - accuracy: 0.5707 - val_loss: 0.9394 - val_accuracy: 0.5357
Epoch 4/150
42/42 [==============================] - 0s 2ms/step - loss: 0.9000 - accuracy: 0.5580 - val_loss: 1.0492 - val_accuracy: 0.5179
Epoch 5/150
42/42 [==============================] - 0s 2ms/step - loss: 0.8955 - accuracy: 0.6010 - val_loss: 1.2346 - val_accuracy: 0.5268
Epoch 6/150
42/42 [==============================] - 0s 2ms/step - loss: 0.8872 - accuracy: 0.5898 - val_loss: 0.9510 - val_accuracy: 0.5804
Epoch 7/150
42/42 [==============================] - 0s 2ms/step - loss: 0.8738 - accuracy: 0.6200 - val_loss: 0.9872 - val_accuracy: 0.5268
Epoch 8/150
42/42 [==============================] - 0s 2ms/step - loss: 0.8387 - accuracy: 0.5946 - val_loss: 1.1425 - val_accuracy: 0.5089
Epoch 9/150
42/42 [==============================] - 0s 2ms/step - loss: 0.8515 - accuracy: 0.6169 - val_loss: 0.9293 - val_accuracy: 0.5446
Epoch 10/150
42/42 [==============================] - 0s 2ms/step - loss: 0.9730 - accuracy: 0.5596 - val_loss: 1.5070 - val_accuracy: 0.5268
Epoch 11/150
42/42 [==============================] - 0s 2ms/step - loss: 0.9004 - accuracy: 0.5628 - val_loss: 0.9662 - val_accuracy: 0.4643
Epoch 12/150
42/42 [==============================] - 0s 2ms/step - loss: 0.7779 - accuracy: 0.6343 - val_loss: 1.0216 - val_accuracy: 0.5268
Epoch 13/150
42/42 [==============================] - 0s 2ms/step - loss: 0.8505 - accuracy: 0.6010 - val_loss: 0.8268 - val_accuracy: 0.6339
Epoch 14/150
42/42 [==============================] - 0s 2ms/step - loss: 0.8515 - accuracy: 0.6359 - val_loss: 0.9324 - val_accuracy: 0.5893
Epoch 15/150
42/42 [==============================] - 0s 2ms/step - loss: 0.8051 - accuracy: 0.6439 - val_loss: 0.8081 - val_accuracy: 0.7054
Epoch 16/150
42/42 [==============================] - 0s 2ms/step - loss: 0.7560 - accuracy: 0.6486 - val_loss: 1.0065 - val_accuracy: 0.5000
Epoch 17/150
42/42 [==============================] - 0s 3ms/step - loss: 0.7378 - accuracy: 0.6486 - val_loss: 0.7666 - val_accuracy: 0.6161
Epoch 18/150
42/42 [==============================] - 0s 2ms/step - loss: 0.7705 - accuracy: 0.6248 - val_loss: 0.8087 - val_accuracy: 0.6250
Epoch 19/150
42/42 [==============================] - 0s 2ms/step - loss: 0.7694 - accuracy: 0.6232 - val_loss: 0.8089 - val_accuracy: 0.6071
Epoch 20/150
42/42 [==============================] - 0s 2ms/step - loss: 0.7206 - accuracy: 0.6677 - val_loss: 0.7415 - val_accuracy: 0.6339
Epoch 21/150
42/42 [==============================] - 0s 2ms/step - loss: 0.7482 - accuracy: 0.6343 - val_loss: 0.7473 - val_accuracy: 0.6071
Epoch 22/150
42/42 [==============================] - 0s 2ms/step - loss: 0.7041 - accuracy: 0.6582 - val_loss: 0.7307 - val_accuracy: 0.6696
Epoch 23/150
42/42 [==============================] - 0s 2ms/step - loss: 0.7422 - accuracy: 0.6518 - val_loss: 0.7760 - val_accuracy: 0.5804
Epoch 24/150
42/42 [==============================] - 0s 2ms/step - loss: 0.6905 - accuracy: 0.6598 - val_loss: 0.7346 - val_accuracy: 0.5982
Epoch 25/150
42/42 [==============================] - 0s 2ms/step - loss: 0.6967 - accuracy: 0.6598 - val_loss: 0.7591 - val_accuracy: 0.6429
Epoch 26/150
42/42 [==============================] - 0s 2ms/step - loss: 0.6898 - accuracy: 0.6582 - val_loss: 0.7813 - val_accuracy: 0.6161
Epoch 27/150
42/42 [==============================] - 0s 2ms/step - loss: 0.7042 - accuracy: 0.6757 - val_loss: 0.7798 - val_accuracy: 0.6339
Epoch 28/150
42/42 [==============================] - 0s 2ms/step - loss: 0.7042 - accuracy: 0.6645 - val_loss: 0.8256 - val_accuracy: 0.5982
Epoch 29/150
42/42 [==============================] - 0s 2ms/step - loss: 0.7058 - accuracy: 0.6614 - val_loss: 0.7130 - val_accuracy: 0.6250
Epoch 30/150
42/42 [==============================] - 0s 2ms/step - loss: 0.6862 - accuracy: 0.6677 - val_loss: 0.7204 - val_accuracy: 0.7321
Epoch 31/150
42/42 [==============================] - 0s 2ms/step - loss: 0.6450 - accuracy: 0.6900 - val_loss: 0.7818 - val_accuracy: 0.5982
Epoch 32/150
42/42 [==============================] - 0s 2ms/step - loss: 0.8793 - accuracy: 0.5978 - val_loss: 0.7702 - val_accuracy: 0.6518
Epoch 33/150
42/42 [==============================] - 0s 2ms/step - loss: 0.7374 - accuracy: 0.6471 - val_loss: 0.6871 - val_accuracy: 0.6964
Epoch 34/150
42/42 [==============================] - 0s 2ms/step - loss: 0.6574 - accuracy: 0.6932 - val_loss: 0.8544 - val_accuracy: 0.6071
Epoch 35/150
42/42 [==============================] - 0s 2ms/step - loss: 0.6777 - accuracy: 0.6836 - val_loss: 0.7507 - val_accuracy: 0.5804
Epoch 36/150
42/42 [==============================] - 0s 2ms/step - loss: 0.6705 - accuracy: 0.7027 - val_loss: 0.6849 - val_accuracy: 0.6429
Epoch 37/150
42/42 [==============================] - 0s 2ms/step - loss: 0.6831 - accuracy: 0.6645 - val_loss: 0.6861 - val_accuracy: 0.5982
Epoch 38/150
42/42 [==============================] - 0s 2ms/step - loss: 0.6284 - accuracy: 0.6995 - val_loss: 0.6885 - val_accuracy: 0.5982
Epoch 39/150
42/42 [==============================] - 0s 3ms/step - loss: 0.7094 - accuracy: 0.6614 - val_loss: 0.6589 - val_accuracy: 0.6429
Epoch 40/150
42/42 [==============================] - 0s 2ms/step - loss: 0.6959 - accuracy: 0.6757 - val_loss: 0.7850 - val_accuracy: 0.5982
Epoch 41/150
42/42 [==============================] - 0s 2ms/step - loss: 0.6377 - accuracy: 0.7107 - val_loss: 0.6601 - val_accuracy: 0.6964
Epoch 42/150
42/42 [==============================] - 0s 2ms/step - loss: 0.6109 - accuracy: 0.7170 - val_loss: 0.6713 - val_accuracy: 0.6875
Epoch 43/150
42/42 [==============================] - 0s 2ms/step - loss: 0.6007 - accuracy: 0.7393 - val_loss: 0.6932 - val_accuracy: 0.7054
Epoch 44/150
42/42 [==============================] - 0s 2ms/step - loss: 0.6737 - accuracy: 0.6963 - val_loss: 0.7397 - val_accuracy: 0.7143
Epoch 45/150
42/42 [==============================] - 0s 2ms/step - loss: 0.6859 - accuracy: 0.6979 - val_loss: 0.7514 - val_accuracy: 0.6429
Epoch 46/150
42/42 [==============================] - 0s 2ms/step - loss: 0.6168 - accuracy: 0.7186 - val_loss: 0.8445 - val_accuracy: 0.6161
Epoch 47/150
42/42 [==============================] - 0s 2ms/step - loss: 0.7418 - accuracy: 0.6677 - val_loss: 0.7712 - val_accuracy: 0.5893
Epoch 48/150
42/42 [==============================] - 0s 2ms/step - loss: 0.6801 - accuracy: 0.7075 - val_loss: 0.6722 - val_accuracy: 0.7411
Epoch 49/150
42/42 [==============================] - 0s 2ms/step - loss: 0.5877 - accuracy: 0.7393 - val_loss: 0.6783 - val_accuracy: 0.6696
Epoch 50/150
42/42 [==============================] - 0s 2ms/step - loss: 0.6121 - accuracy: 0.7281 - val_loss: 0.6274 - val_accuracy: 0.7321
Epoch 51/150
42/42 [==============================] - 0s 2ms/step - loss: 0.6521 - accuracy: 0.6963 - val_loss: 0.6300 - val_accuracy: 0.7411
Epoch 52/150
42/42 [==============================] - 0s 2ms/step - loss: 0.5880 - accuracy: 0.7504 - val_loss: 0.6456 - val_accuracy: 0.6786
Epoch 53/150
42/42 [==============================] - 0s 2ms/step - loss: 0.6312 - accuracy: 0.7266 - val_loss: 0.8025 - val_accuracy: 0.5536
Epoch 54/150
42/42 [==============================] - 0s 2ms/step - loss: 0.6132 - accuracy: 0.7027 - val_loss: 0.7209 - val_accuracy: 0.6518
Epoch 55/150
42/42 [==============================] - 0s 2ms/step - loss: 0.6634 - accuracy: 0.6852 - val_loss: 0.7073 - val_accuracy: 0.6875
Epoch 56/150
42/42 [==============================] - 0s 2ms/step - loss: 0.6428 - accuracy: 0.7043 - val_loss: 0.6417 - val_accuracy: 0.6696
Epoch 57/150
42/42 [==============================] - 0s 2ms/step - loss: 0.6140 - accuracy: 0.7154 - val_loss: 0.6690 - val_accuracy: 0.6607
Epoch 58/150
42/42 [==============================] - 0s 2ms/step - loss: 0.5779 - accuracy: 0.7504 - val_loss: 0.6394 - val_accuracy: 0.6339
Epoch 59/150
42/42 [==============================] - 0s 2ms/step - loss: 0.5883 - accuracy: 0.7424 - val_loss: 0.5840 - val_accuracy: 0.7143
Epoch 60/150
42/42 [==============================] - 0s 2ms/step - loss: 0.5714 - accuracy: 0.7631 - val_loss: 0.8597 - val_accuracy: 0.6071
Epoch 61/150
42/42 [==============================] - 0s 2ms/step - loss: 0.6060 - accuracy: 0.7154 - val_loss: 0.9389 - val_accuracy: 0.5893
Epoch 62/150
42/42 [==============================] - 0s 2ms/step - loss: 0.6477 - accuracy: 0.7043 - val_loss: 0.6029 - val_accuracy: 0.6607
Epoch 63/150
42/42 [==============================] - 0s 2ms/step - loss: 0.5634 - accuracy: 0.7504 - val_loss: 0.6315 - val_accuracy: 0.6607
Epoch 64/150
42/42 [==============================] - 0s 2ms/step - loss: 0.5908 - accuracy: 0.7488 - val_loss: 0.6130 - val_accuracy: 0.7143
Epoch 65/150
42/42 [==============================] - 0s 2ms/step - loss: 0.5618 - accuracy: 0.7456 - val_loss: 0.5798 - val_accuracy: 0.7411
Epoch 66/150
42/42 [==============================] - 0s 2ms/step - loss: 0.5465 - accuracy: 0.7663 - val_loss: 0.6004 - val_accuracy: 0.7054
Epoch 67/150
42/42 [==============================] - 0s 2ms/step - loss: 0.6057 - accuracy: 0.7377 - val_loss: 0.6292 - val_accuracy: 0.7054
Epoch 68/150
42/42 [==============================] - 0s 2ms/step - loss: 0.6425 - accuracy: 0.7170 - val_loss: 0.5640 - val_accuracy: 0.7321
Epoch 69/150
42/42 [==============================] - 0s 2ms/step - loss: 0.6000 - accuracy: 0.7154 - val_loss: 0.5606 - val_accuracy: 0.7679
Epoch 70/150
42/42 [==============================] - 0s 2ms/step - loss: 0.5665 - accuracy: 0.7424 - val_loss: 0.5839 - val_accuracy: 0.7411
Epoch 71/150
42/42 [==============================] - 0s 2ms/step - loss: 0.5709 - accuracy: 0.7361 - val_loss: 0.5529 - val_accuracy: 0.7679
Epoch 72/150
42/42 [==============================] - 0s 2ms/step - loss: 0.5358 - accuracy: 0.7599 - val_loss: 0.7237 - val_accuracy: 0.7321
Epoch 73/150
42/42 [==============================] - 0s 2ms/step - loss: 0.5880 - accuracy: 0.7361 - val_loss: 0.5803 - val_accuracy: 0.7232
Epoch 74/150
42/42 [==============================] - 0s 2ms/step - loss: 0.5458 - accuracy: 0.7711 - val_loss: 0.5434 - val_accuracy: 0.7768
Epoch 75/150
42/42 [==============================] - 0s 2ms/step - loss: 0.5268 - accuracy: 0.7901 - val_loss: 0.5363 - val_accuracy: 0.7768
Epoch 76/150
42/42 [==============================] - 0s 2ms/step - loss: 0.5268 - accuracy: 0.7886 - val_loss: 0.5722 - val_accuracy: 0.7679
Epoch 77/150
42/42 [==============================] - 0s 2ms/step - loss: 0.6105 - accuracy: 0.7281 - val_loss: 0.5493 - val_accuracy: 0.8036
Epoch 78/150
42/42 [==============================] - 0s 2ms/step - loss: 0.5371 - accuracy: 0.7870 - val_loss: 0.5229 - val_accuracy: 0.7857
Epoch 79/150
42/42 [==============================] - 0s 2ms/step - loss: 0.5111 - accuracy: 0.7997 - val_loss: 0.5315 - val_accuracy: 0.8036
Epoch 80/150
42/42 [==============================] - 0s 2ms/step - loss: 0.5440 - accuracy: 0.7504 - val_loss: 0.5895 - val_accuracy: 0.7321
Epoch 81/150
42/42 [==============================] - 0s 2ms/step - loss: 0.5563 - accuracy: 0.7583 - val_loss: 0.5725 - val_accuracy: 0.7857
Epoch 82/150
42/42 [==============================] - 0s 2ms/step - loss: 0.5046 - accuracy: 0.8076 - val_loss: 0.5408 - val_accuracy: 0.7589
Epoch 83/150
42/42 [==============================] - 0s 2ms/step - loss: 0.5824 - accuracy: 0.7393 - val_loss: 0.8550 - val_accuracy: 0.5625
Epoch 84/150
42/42 [==============================] - 0s 2ms/step - loss: 0.5392 - accuracy: 0.7488 - val_loss: 0.5254 - val_accuracy: 0.8125
Epoch 85/150
42/42 [==============================] - 0s 2ms/step - loss: 0.5254 - accuracy: 0.7758 - val_loss: 0.5193 - val_accuracy: 0.8304
Epoch 86/150
42/42 [==============================] - 0s 2ms/step - loss: 0.5182 - accuracy: 0.7742 - val_loss: 0.5135 - val_accuracy: 0.7679
Epoch 87/150
42/42 [==============================] - 0s 2ms/step - loss: 0.4788 - accuracy: 0.8045 - val_loss: 0.5494 - val_accuracy: 0.7589
Epoch 88/150
42/42 [==============================] - 0s 2ms/step - loss: 0.4961 - accuracy: 0.7981 - val_loss: 0.7635 - val_accuracy: 0.5982
Epoch 89/150
42/42 [==============================] - 0s 2ms/step - loss: 0.5295 - accuracy: 0.7647 - val_loss: 0.6529 - val_accuracy: 0.6429
Epoch 90/150
42/42 [==============================] - 0s 2ms/step - loss: 0.5216 - accuracy: 0.7536 - val_loss: 0.5021 - val_accuracy: 0.7857
Epoch 91/150
42/42 [==============================] - 0s 2ms/step - loss: 0.4741 - accuracy: 0.8140 - val_loss: 0.4889 - val_accuracy: 0.7857
Epoch 92/150
42/42 [==============================] - 0s 2ms/step - loss: 0.5612 - accuracy: 0.7440 - val_loss: 0.7228 - val_accuracy: 0.7143
Epoch 93/150
42/42 [==============================] - 0s 2ms/step - loss: 0.5770 - accuracy: 0.7409 - val_loss: 0.6267 - val_accuracy: 0.6786
Epoch 94/150
42/42 [==============================] - 0s 2ms/step - loss: 0.5631 - accuracy: 0.7520 - val_loss: 0.5478 - val_accuracy: 0.7411
Epoch 95/150
42/42 [==============================] - 0s 2ms/step - loss: 0.5415 - accuracy: 0.7695 - val_loss: 0.7320 - val_accuracy: 0.6339
Epoch 96/150
42/42 [==============================] - 0s 2ms/step - loss: 0.5919 - accuracy: 0.7297 - val_loss: 0.6597 - val_accuracy: 0.6607
Epoch 97/150
42/42 [==============================] - 0s 2ms/step - loss: 0.5361 - accuracy: 0.7583 - val_loss: 0.5364 - val_accuracy: 0.7500
Epoch 98/150
42/42 [==============================] - 0s 2ms/step - loss: 0.5128 - accuracy: 0.7981 - val_loss: 0.5289 - val_accuracy: 0.7679
Epoch 99/150
42/42 [==============================] - 0s 2ms/step - loss: 0.4674 - accuracy: 0.8092 - val_loss: 0.4932 - val_accuracy: 0.8304
Epoch 100/150
42/42 [==============================] - 0s 2ms/step - loss: 0.5007 - accuracy: 0.7711 - val_loss: 0.4961 - val_accuracy: 0.7679
Epoch 101/150
42/42 [==============================] - 0s 2ms/step - loss: 0.4755 - accuracy: 0.8076 - val_loss: 0.4810 - val_accuracy: 0.8304
Epoch 102/150
42/42 [==============================] - 0s 2ms/step - loss: 0.5066 - accuracy: 0.7901 - val_loss: 0.4930 - val_accuracy: 0.8214
Epoch 103/150
42/42 [==============================] - 0s 2ms/step - loss: 0.4671 - accuracy: 0.8172 - val_loss: 0.5213 - val_accuracy: 0.7500
Epoch 104/150
42/42 [==============================] - 0s 2ms/step - loss: 0.4467 - accuracy: 0.8108 - val_loss: 0.4751 - val_accuracy: 0.8036
Epoch 105/150
42/42 [==============================] - 0s 2ms/step - loss: 0.4971 - accuracy: 0.7838 - val_loss: 0.5881 - val_accuracy: 0.7321
Epoch 106/150
42/42 [==============================] - 0s 2ms/step - loss: 0.5056 - accuracy: 0.7742 - val_loss: 0.4593 - val_accuracy: 0.8393
Epoch 107/150
42/42 [==============================] - 0s 2ms/step - loss: 0.4817 - accuracy: 0.8172 - val_loss: 0.7131 - val_accuracy: 0.6518
Epoch 108/150
42/42 [==============================] - 0s 2ms/step - loss: 0.4726 - accuracy: 0.8013 - val_loss: 0.5633 - val_accuracy: 0.7321
Epoch 109/150
42/42 [==============================] - 0s 2ms/step - loss: 0.4841 - accuracy: 0.7822 - val_loss: 0.4582 - val_accuracy: 0.8393
Epoch 110/150
42/42 [==============================] - 0s 2ms/step - loss: 0.4692 - accuracy: 0.8092 - val_loss: 0.5228 - val_accuracy: 0.7500
Epoch 111/150
42/42 [==============================] - 0s 2ms/step - loss: 0.4478 - accuracy: 0.8108 - val_loss: 0.4706 - val_accuracy: 0.7768
Epoch 112/150
42/42 [==============================] - 0s 2ms/step - loss: 0.4879 - accuracy: 0.7933 - val_loss: 0.5371 - val_accuracy: 0.7589
Epoch 113/150
42/42 [==============================] - 0s 2ms/step - loss: 0.4386 - accuracy: 0.8235 - val_loss: 0.4984 - val_accuracy: 0.7500
Epoch 114/150
42/42 [==============================] - 0s 2ms/step - loss: 0.4623 - accuracy: 0.7838 - val_loss: 0.5855 - val_accuracy: 0.7143
Epoch 115/150
42/42 [==============================] - 0s 2ms/step - loss: 0.4805 - accuracy: 0.7965 - val_loss: 0.4710 - val_accuracy: 0.8214
Epoch 116/150
42/42 [==============================] - 0s 2ms/step - loss: 0.4557 - accuracy: 0.8203 - val_loss: 0.4289 - val_accuracy: 0.8393
Epoch 117/150
42/42 [==============================] - 0s 2ms/step - loss: 0.4149 - accuracy: 0.8426 - val_loss: 0.4573 - val_accuracy: 0.8125
Epoch 118/150
42/42 [==============================] - 0s 2ms/step - loss: 0.4198 - accuracy: 0.8426 - val_loss: 0.5044 - val_accuracy: 0.8214
Epoch 119/150
42/42 [==============================] - 0s 2ms/step - loss: 0.4064 - accuracy: 0.8426 - val_loss: 0.4645 - val_accuracy: 0.8571
Epoch 120/150
42/42 [==============================] - 0s 2ms/step - loss: 0.4332 - accuracy: 0.8108 - val_loss: 0.4384 - val_accuracy: 0.8125
Epoch 121/150
42/42 [==============================] - 0s 2ms/step - loss: 0.4244 - accuracy: 0.8172 - val_loss: 0.4538 - val_accuracy: 0.8393
Epoch 122/150
42/42 [==============================] - 0s 2ms/step - loss: 0.5670 - accuracy: 0.7536 - val_loss: 0.5498 - val_accuracy: 0.7411
Epoch 123/150
42/42 [==============================] - 0s 2ms/step - loss: 0.6258 - accuracy: 0.7266 - val_loss: 0.4367 - val_accuracy: 0.8393
Epoch 124/150
42/42 [==============================] - 0s 2ms/step - loss: 0.4532 - accuracy: 0.8060 - val_loss: 0.4494 - val_accuracy: 0.8125
Epoch 125/150
42/42 [==============================] - 0s 2ms/step - loss: 0.4330 - accuracy: 0.8156 - val_loss: 0.7741 - val_accuracy: 0.6429
Epoch 126/150
42/42 [==============================] - 0s 2ms/step - loss: 0.5057 - accuracy: 0.7774 - val_loss: 0.4527 - val_accuracy: 0.8036
Epoch 127/150
42/42 [==============================] - 0s 3ms/step - loss: 0.4669 - accuracy: 0.8108 - val_loss: 0.4691 - val_accuracy: 0.8036
Epoch 128/150
42/42 [==============================] - 0s 2ms/step - loss: 0.4287 - accuracy: 0.8267 - val_loss: 0.4189 - val_accuracy: 0.8214
Epoch 129/150
42/42 [==============================] - 0s 2ms/step - loss: 0.4015 - accuracy: 0.8315 - val_loss: 0.5260 - val_accuracy: 0.7500
Epoch 130/150
42/42 [==============================] - 0s 2ms/step - loss: 0.4428 - accuracy: 0.8013 - val_loss: 0.5850 - val_accuracy: 0.6786
Epoch 131/150
42/42 [==============================] - 0s 2ms/step - loss: 0.4305 - accuracy: 0.8315 - val_loss: 0.5077 - val_accuracy: 0.7768
Epoch 132/150
42/42 [==============================] - 0s 2ms/step - loss: 0.4084 - accuracy: 0.8267 - val_loss: 0.4328 - val_accuracy: 0.8214
Epoch 133/150
42/42 [==============================] - 0s 2ms/step - loss: 0.3988 - accuracy: 0.8490 - val_loss: 0.4997 - val_accuracy: 0.7857
Epoch 134/150
42/42 [==============================] - 0s 2ms/step - loss: 0.4544 - accuracy: 0.8013 - val_loss: 0.4098 - val_accuracy: 0.8125
Epoch 135/150
42/42 [==============================] - 0s 2ms/step - loss: 0.3988 - accuracy: 0.8378 - val_loss: 0.4672 - val_accuracy: 0.8571
Epoch 136/150
42/42 [==============================] - 0s 2ms/step - loss: 0.3994 - accuracy: 0.8283 - val_loss: 0.5769 - val_accuracy: 0.7589
Epoch 137/150
42/42 [==============================] - 0s 3ms/step - loss: 0.4145 - accuracy: 0.8156 - val_loss: 0.4016 - val_accuracy: 0.8393
Epoch 138/150
42/42 [==============================] - 0s 2ms/step - loss: 0.3824 - accuracy: 0.8506 - val_loss: 0.5700 - val_accuracy: 0.7857
Epoch 139/150
42/42 [==============================] - 0s 2ms/step - loss: 0.3997 - accuracy: 0.8378 - val_loss: 0.4817 - val_accuracy: 0.7679
Epoch 140/150
42/42 [==============================] - 0s 2ms/step - loss: 0.3783 - accuracy: 0.8490 - val_loss: 0.4410 - val_accuracy: 0.8214
Epoch 141/150
42/42 [==============================] - 0s 2ms/step - loss: 0.3908 - accuracy: 0.8188 - val_loss: 0.5193 - val_accuracy: 0.7679
Epoch 142/150
42/42 [==============================] - 0s 2ms/step - loss: 0.4194 - accuracy: 0.8172 - val_loss: 0.4430 - val_accuracy: 0.7679
Epoch 143/150
42/42 [==============================] - 0s 2ms/step - loss: 0.4135 - accuracy: 0.8219 - val_loss: 0.4415 - val_accuracy: 0.7679
Epoch 144/150
42/42 [==============================] - 0s 2ms/step - loss: 0.4347 - accuracy: 0.8045 - val_loss: 0.5366 - val_accuracy: 0.7857
Epoch 145/150
42/42 [==============================] - 0s 2ms/step - loss: 0.4073 - accuracy: 0.8315 - val_loss: 0.4410 - val_accuracy: 0.8214
Epoch 146/150
42/42 [==============================] - 0s 2ms/step - loss: 0.4026 - accuracy: 0.8394 - val_loss: 0.4013 - val_accuracy: 0.8661
Epoch 147/150
42/42 [==============================] - 0s 2ms/step - loss: 0.3793 - accuracy: 0.8458 - val_loss: 0.5085 - val_accuracy: 0.7946
Epoch 148/150
42/42 [==============================] - 0s 2ms/step - loss: 0.4227 - accuracy: 0.8347 - val_loss: 0.4853 - val_accuracy: 0.8214
Epoch 149/150
42/42 [==============================] - 0s 2ms/step - loss: 0.4286 - accuracy: 0.8235 - val_loss: 0.4343 - val_accuracy: 0.7946
Epoch 150/150
42/42 [==============================] - 0s 2ms/step - loss: 0.3879 - accuracy: 0.8410 - val_loss: 0.5773 - val_accuracy: 0.7768
10/10 [==============================] - 0s 999us/step - loss: 0.4838 - accuracy: 0.7704
accuracy: 77.0440%
In [35]:
def train_eval_model_2(x1, y1, x2, y2, n1, n2, n):
    """Función para realizar el entrenamiento de la red neuronal. Asimismo,
    se realiza la evaluación del mismo.
    Parámetros:
    * x1: X_train.
    * y1: y_train.
    * x2: X_test.
    * y2: y_test.
    * n1: número de neuronas en la primer capa oculta.
    * n2: número de neuronas en la segunda capa oculta.
    * n: número de épocas.
    * k: batch_size."""
    # Instanciamos el modelo secuencial
    model = Sequential()
    # Capa de entrada y primera capa oculta 
    model.add(Dense(n1, activation='relu', input_shape=(7,)))
    # Capa oculta 2
    model.add(Dense(n2, activation='relu'))
    # Capa de salida
    model.add(Dense(3, activation='softmax'))
    # Copilacion del modelo
    model.compile(loss='categorical_crossentropy', optimizer='adam', metrics=['accuracy'])
    # Entrenamiento del modelo.
    # Destinamos el 15% de los datos de entrenamiento para la validacion 
    history = model.fit(x1, y1, epochs=n, validation_split=0.15)
    # Evalucacion del modelo
    scores = model.evaluate(x2, y2)
    print('%s: %.4f%%' % (model.metrics_names[1], scores[1] * 100))
    # Grafico de la precision del entrenamiento vs precision de la validacion
    plt.plot(history.history['accuracy'])
    plt.plot(history.history['val_accuracy'])
    plt.title('Precisión del modelo')
    plt.xlabel('Epoch')
    plt.ylabel('Precisión')
    plt.legend(['Train', 'Validación'])
    plt.show()
In [36]:
train_eval_model_2(X_train, y_train, X_test, y_test, 55, 30, 50)

Epoch 1/50
20/20 [==============================] - 1s 12ms/step - loss: 13.9900 - accuracy: 0.4324 - val_loss: 1.7509 - val_accuracy: 0.4554
Epoch 2/50
20/20 [==============================] - 0s 3ms/step - loss: 2.0137 - accuracy: 0.4356 - val_loss: 1.3986 - val_accuracy: 0.4554
Epoch 3/50
20/20 [==============================] - 0s 3ms/step - loss: 1.0370 - accuracy: 0.5326 - val_loss: 1.0501 - val_accuracy: 0.5714
Epoch 4/50
20/20 [==============================] - 0s 3ms/step - loss: 0.9475 - accuracy: 0.5040 - val_loss: 1.0329 - val_accuracy: 0.5625
Epoch 5/50
20/20 [==============================] - 0s 3ms/step - loss: 0.9353 - accuracy: 0.5310 - val_loss: 1.0442 - val_accuracy: 0.4732
Epoch 6/50
20/20 [==============================] - 0s 3ms/step - loss: 0.9629 - accuracy: 0.5882 - val_loss: 1.0296 - val_accuracy: 0.5089
Epoch 7/50
20/20 [==============================] - 0s 3ms/step - loss: 0.8972 - accuracy: 0.5660 - val_loss: 1.0180 - val_accuracy: 0.4196
Epoch 8/50
20/20 [==============================] - 0s 3ms/step - loss: 0.8920 - accuracy: 0.5453 - val_loss: 0.9817 - val_accuracy: 0.5357
Epoch 9/50
20/20 [==============================] - 0s 4ms/step - loss: 0.8891 - accuracy: 0.6169 - val_loss: 0.9729 - val_accuracy: 0.5446
Epoch 10/50
20/20 [==============================] - 0s 3ms/step - loss: 0.9232 - accuracy: 0.5644 - val_loss: 1.0491 - val_accuracy: 0.5893
Epoch 11/50
20/20 [==============================] - 0s 4ms/step - loss: 0.8816 - accuracy: 0.5898 - val_loss: 0.9585 - val_accuracy: 0.5714
Epoch 12/50
20/20 [==============================] - 0s 3ms/step - loss: 0.8319 - accuracy: 0.6121 - val_loss: 1.0043 - val_accuracy: 0.7411
Epoch 13/50
20/20 [==============================] - 0s 3ms/step - loss: 0.8580 - accuracy: 0.5739 - val_loss: 0.9523 - val_accuracy: 0.5357
Epoch 14/50
20/20 [==============================] - 0s 3ms/step - loss: 0.8705 - accuracy: 0.5739 - val_loss: 0.9486 - val_accuracy: 0.4732
Epoch 15/50
20/20 [==============================] - 0s 4ms/step - loss: 0.8285 - accuracy: 0.6025 - val_loss: 0.8891 - val_accuracy: 0.6339
Epoch 16/50
20/20 [==============================] - 0s 3ms/step - loss: 0.7847 - accuracy: 0.6471 - val_loss: 0.8775 - val_accuracy: 0.6339
Epoch 17/50
20/20 [==============================] - 0s 3ms/step - loss: 0.7952 - accuracy: 0.6010 - val_loss: 0.8938 - val_accuracy: 0.6071
Epoch 18/50
20/20 [==============================] - 0s 3ms/step - loss: 0.9453 - accuracy: 0.5612 - val_loss: 0.8804 - val_accuracy: 0.6429
Epoch 19/50
20/20 [==============================] - 0s 3ms/step - loss: 0.8910 - accuracy: 0.5930 - val_loss: 0.8523 - val_accuracy: 0.6607
Epoch 20/50
20/20 [==============================] - 0s 3ms/step - loss: 0.7737 - accuracy: 0.6614 - val_loss: 1.0001 - val_accuracy: 0.5357
Epoch 21/50
20/20 [==============================] - 0s 3ms/step - loss: 0.8252 - accuracy: 0.6184 - val_loss: 0.8519 - val_accuracy: 0.5982
Epoch 22/50
20/20 [==============================] - 0s 3ms/step - loss: 0.7756 - accuracy: 0.6518 - val_loss: 0.8990 - val_accuracy: 0.5804
Epoch 23/50
20/20 [==============================] - 0s 3ms/step - loss: 0.7292 - accuracy: 0.6630 - val_loss: 0.9244 - val_accuracy: 0.5804
Epoch 24/50
20/20 [==============================] - 0s 3ms/step - loss: 0.7804 - accuracy: 0.6534 - val_loss: 0.8490 - val_accuracy: 0.5982
Epoch 25/50
20/20 [==============================] - 0s 3ms/step - loss: 0.7038 - accuracy: 0.6900 - val_loss: 0.7875 - val_accuracy: 0.6786
Epoch 26/50
20/20 [==============================] - 0s 3ms/step - loss: 0.7197 - accuracy: 0.6932 - val_loss: 0.7820 - val_accuracy: 0.6607
Epoch 27/50
20/20 [==============================] - 0s 3ms/step - loss: 0.7119 - accuracy: 0.6757 - val_loss: 0.7811 - val_accuracy: 0.6964
Epoch 28/50
20/20 [==============================] - 0s 3ms/step - loss: 0.7007 - accuracy: 0.6900 - val_loss: 0.9782 - val_accuracy: 0.6161
Epoch 29/50
20/20 [==============================] - 0s 3ms/step - loss: 0.7207 - accuracy: 0.6709 - val_loss: 0.8066 - val_accuracy: 0.6696
Epoch 30/50
20/20 [==============================] - 0s 3ms/step - loss: 0.6713 - accuracy: 0.7202 - val_loss: 0.7609 - val_accuracy: 0.7232
Epoch 31/50
20/20 [==============================] - 0s 3ms/step - loss: 0.6945 - accuracy: 0.7154 - val_loss: 0.7744 - val_accuracy: 0.6161
Epoch 32/50
20/20 [==============================] - 0s 3ms/step - loss: 0.6479 - accuracy: 0.7361 - val_loss: 0.7649 - val_accuracy: 0.6518
Epoch 33/50
20/20 [==============================] - 0s 3ms/step - loss: 0.6417 - accuracy: 0.7424 - val_loss: 0.7904 - val_accuracy: 0.6964
Epoch 34/50
20/20 [==============================] - 0s 3ms/step - loss: 0.6375 - accuracy: 0.7377 - val_loss: 0.7164 - val_accuracy: 0.6964
Epoch 35/50
20/20 [==============================] - 0s 3ms/step - loss: 0.6823 - accuracy: 0.7011 - val_loss: 0.7212 - val_accuracy: 0.7500
Epoch 36/50
20/20 [==============================] - 0s 3ms/step - loss: 0.6411 - accuracy: 0.7345 - val_loss: 0.7957 - val_accuracy: 0.6429
Epoch 37/50
20/20 [==============================] - 0s 3ms/step - loss: 0.6638 - accuracy: 0.7043 - val_loss: 0.7528 - val_accuracy: 0.6607
Epoch 38/50
20/20 [==============================] - 0s 3ms/step - loss: 0.7517 - accuracy: 0.6661 - val_loss: 1.0432 - val_accuracy: 0.6161
Epoch 39/50
20/20 [==============================] - 0s 3ms/step - loss: 0.7115 - accuracy: 0.6963 - val_loss: 0.7995 - val_accuracy: 0.6607
Epoch 40/50
20/20 [==============================] - 0s 3ms/step - loss: 0.7378 - accuracy: 0.6884 - val_loss: 0.7625 - val_accuracy: 0.6786
Epoch 41/50
20/20 [==============================] - 0s 3ms/step - loss: 0.6710 - accuracy: 0.6979 - val_loss: 0.6685 - val_accuracy: 0.7411
Epoch 42/50
20/20 [==============================] - 0s 3ms/step - loss: 0.5998 - accuracy: 0.7536 - val_loss: 0.6730 - val_accuracy: 0.7321
Epoch 43/50
20/20 [==============================] - 0s 3ms/step - loss: 0.6273 - accuracy: 0.7138 - val_loss: 0.6488 - val_accuracy: 0.7500
Epoch 44/50
20/20 [==============================] - 0s 3ms/step - loss: 0.6136 - accuracy: 0.7377 - val_loss: 0.7308 - val_accuracy: 0.6607
Epoch 45/50
20/20 [==============================] - 0s 3ms/step - loss: 0.5797 - accuracy: 0.7631 - val_loss: 0.6640 - val_accuracy: 0.7768
Epoch 46/50
20/20 [==============================] - 0s 3ms/step - loss: 0.6055 - accuracy: 0.7329 - val_loss: 0.7130 - val_accuracy: 0.6518
Epoch 47/50
20/20 [==============================] - 0s 3ms/step - loss: 0.6008 - accuracy: 0.7695 - val_loss: 0.6252 - val_accuracy: 0.7679
Epoch 48/50
20/20 [==============================] - 0s 3ms/step - loss: 0.5672 - accuracy: 0.7917 - val_loss: 0.6275 - val_accuracy: 0.7321
Epoch 49/50
20/20 [==============================] - 0s 3ms/step - loss: 0.5455 - accuracy: 0.7901 - val_loss: 0.6077 - val_accuracy: 0.7857
Epoch 50/50
20/20 [==============================] - 0s 3ms/step - loss: 0.5331 - accuracy: 0.7901 - val_loss: 0.6720 - val_accuracy: 0.7946
10/10 [==============================] - 0s 1ms/step - loss: 0.5806 - accuracy: 0.7925
accuracy: 79.2453%
In [40]:
from sklearn.ensemble import RandomForestClassifier

# Configuramos
# Tomamos 20 arboles.
# Criterio a considerar: gini.
b_aleatorio = RandomForestClassifier(n_estimators=20, criterion='gini', max_depth=None, min_samples_split=2)

# Entrenamiento
model = b_aleatorio.fit(X_train, y_train)

# METRICAS
# Hacemos las predicciones
y_pred = model.predict(X_test)
print(f"f1 score: {f1_score(y_test, y_pred, average='weighted')}")
print(f"Recall score: {recall_score(y_test, y_pred, average='weighted')}")
print(f"precision score: {precision_score(y_test, y_pred, average='weighted')}")
print(f"accuracy score: {accuracy_score(y_test, y_pred)}")

f1 score: 0.9984217396463746
Recall score: 0.9968553459119497
precision score: 1.0
accuracy score: 0.9968553459119497
In [46]:
# Creamos un dataframe con los valores de y_test y y_pred
aux_df = pd.DataFrame(enc.inverse_transform(y_pred))
aux_df['y_test'] = enc.inverse_transform(y_test)
aux_df = aux_df.rename(columns={0: 'y_pred'})

# veamos
aux_df
Out[46]:
y_pred y_test
0 medium medium
1 medium medium
2 low low
3 low low
4 medium medium
... ... ...
313 low low
314 low low
315 high high
316 low low
317 low low

318 rows × 2 columns

In [47]:
def clasificador(x):
    if x == 'low':
        return 0
    elif x == 'medium':
        return 1
    else:
        return 2

aux_df['value1'] = aux_df.y_pred.apply(lambda x: clasificador(x))
aux_df['value2'] = aux_df.y_test.apply(lambda x: clasificador(x))

# Creamos una columna nueva donde, si la resta entre las columnas value1 y value2
# es de cero, entonces el clasificador ha predicho bien el resultado, caso contrario
# el clasificador ha errado:
aux_df['correcto'] = aux_df.value1 - aux_df.value2

aux_df
Out[47]:
y_pred y_test value1 value2 correcto
0 medium medium 1 1 0
1 medium medium 1 1 0
2 low low 0 0 0
3 low low 0 0 0
4 medium medium 1 1 0
... ... ... ... ... ...
313 low low 0 0 0
314 low low 0 0 0
315 high high 2 2 0
316 low low 0 0 0
317 low low 0 0 0

318 rows × 5 columns

In [48]:
# Veamos cuantos ceros obtuvimos, es decir, cuantos valores bien predichos
# hemos obtenido con base en el clasificador
aux_df.correcto.value_counts()
Out[48]:
0    317
2      1
Name: correcto, dtype: int64
In [49]:
aux_df[aux_df.correcto != 0]
Out[49]:
y_pred y_test value1 value2 correcto
240 None low 2 0 2