Ejemplo empresas en re organización#
import pandas as pd
import numpy as np
import matplotlib.pyplot as plt
import seaborn as sns
from sklearn.model_selection import train_test_split
from sklearn.preprocessing import StandardScaler
from sklearn.metrics import accuracy_score
from sklearn.metrics import classification_report, confusion_matrix, accuracy_score, recall_score, precision_score
path = "BD empresas en re organización.xlsx"
xls = pd.ExcelFile(path)
df = pd.read_excel(path, sheet_name=xls.sheet_names[0])
df.head()
| Razón Social | Margen EBIT | Carga financiera | Margen neto | CxC | CxP | Solvencia | Apalancamiento | En Reorganización | |
|---|---|---|---|---|---|---|---|---|---|
| 0 | AACER SAS | 0.071690 | 0.000000 | 0.042876 | 0.104095 | 0.153192 | 1.877078 | 1.642505 | 0 |
| 1 | ABARROTES EL ROMPOY SAS | 0.017816 | 0.000000 | 0.010767 | 0.018414 | 0.000000 | 0.000000 | 0.865044 | 0 |
| 2 | ABASTECIMIENTOS INDUSTRIALES SAS | 0.144646 | 0.054226 | 0.059784 | 0.227215 | 0.025591 | 1.077412 | 1.272299 | 0 |
| 3 | ACME LEON PLASTICOS SAS | 0.004465 | 0.000000 | -0.013995 | 0.073186 | 0.127866 | 0.000000 | 1.391645 | 0 |
| 4 | ADVANCED PRODUCTS COLOMBIA SAS | 0.141829 | 0.050810 | 0.053776 | 0.398755 | 0.147678 | 0.675073 | 2.118774 | 0 |
# ------------------------
# Selección de variables
# ------------------------
variables_seleccionadas = ['Margen EBIT',
'Carga financiera',
'Margen neto',
'CxC',
'CxP',
'Solvencia',
'Apalancamiento']
# Variable objetivo
target = 'En Reorganización'
# ------------------------
# Preparar datos
# ------------------------
X = df[variables_seleccionadas]
y = df[target]
# Estandarizar variables
scaler = StandardScaler()
X_scaled = scaler.fit_transform(X)
# Dividir en entrenamiento y prueba (70%-30%)
X_train, X_test, y_train, y_test = train_test_split(X_scaled, y, test_size=0.3, random_state=35, stratify=y)
type(X_train)
numpy.ndarray
X_train.shape
(440, 7)
from keras.models import Sequential
from keras.layers import Dense, Input, Dropout
from keras import optimizers
model = Sequential()
model.add(Input(shape=(X.shape[1],))) # Definimos la forma de entrada
model.add(Dense(units=10, activation="relu")) # Primera capa oculta
model.add(Dense(units=10, activation="relu")) # Segunda capa oculta
model.add(Dense(units=1, activation="sigmoid")) # Capa de salida
model.compile(loss="binary_crossentropy", metrics=["accuracy"], optimizer=optimizers.Adam(learning_rate=0.001))
history = model.fit(X_train, y_train, epochs=100,
validation_data=(X_test, y_test),
batch_size=32,
verbose=1)
Epoch 1/100
[1m14/14[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m3s[0m 56ms/step - accuracy: 0.6487 - loss: 0.6540 - val_accuracy: 0.6667 - val_loss: 0.6418
Epoch 2/100
[1m14/14[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m1s[0m 58ms/step - accuracy: 0.7216 - loss: 0.6285 - val_accuracy: 0.6878 - val_loss: 0.6257
Epoch 3/100
[1m14/14[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m1s[0m 27ms/step - accuracy: 0.7178 - loss: 0.6154 - val_accuracy: 0.6931 - val_loss: 0.6115
Epoch 4/100
[1m14/14[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m1s[0m 26ms/step - accuracy: 0.7566 - loss: 0.5849 - val_accuracy: 0.7090 - val_loss: 0.5973
Epoch 5/100
[1m14/14[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m1s[0m 30ms/step - accuracy: 0.7741 - loss: 0.5819 - val_accuracy: 0.6984 - val_loss: 0.5838
Epoch 6/100
[1m14/14[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m1s[0m 35ms/step - accuracy: 0.7809 - loss: 0.5538 - val_accuracy: 0.7037 - val_loss: 0.5714
Epoch 7/100
[1m14/14[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m1s[0m 56ms/step - accuracy: 0.7816 - loss: 0.5547 - val_accuracy: 0.7196 - val_loss: 0.5600
Epoch 8/100
[1m14/14[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m1s[0m 35ms/step - accuracy: 0.7796 - loss: 0.5158 - val_accuracy: 0.7249 - val_loss: 0.5491
Epoch 9/100
[1m14/14[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m1s[0m 37ms/step - accuracy: 0.7557 - loss: 0.5264 - val_accuracy: 0.7302 - val_loss: 0.5399
Epoch 10/100
[1m14/14[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m1s[0m 45ms/step - accuracy: 0.7519 - loss: 0.5069 - val_accuracy: 0.7460 - val_loss: 0.5315
Epoch 11/100
[1m14/14[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m1s[0m 24ms/step - accuracy: 0.7648 - loss: 0.4940 - val_accuracy: 0.7566 - val_loss: 0.5234
Epoch 12/100
[1m14/14[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m1s[0m 34ms/step - accuracy: 0.7621 - loss: 0.4989 - val_accuracy: 0.7566 - val_loss: 0.5171
Epoch 13/100
[1m14/14[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 14ms/step - accuracy: 0.7717 - loss: 0.4844 - val_accuracy: 0.7672 - val_loss: 0.5120
Epoch 14/100
[1m14/14[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 14ms/step - accuracy: 0.7685 - loss: 0.5119 - val_accuracy: 0.7725 - val_loss: 0.5065
Epoch 15/100
[1m14/14[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 16ms/step - accuracy: 0.7876 - loss: 0.4653 - val_accuracy: 0.7672 - val_loss: 0.5013
Epoch 16/100
[1m14/14[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 17ms/step - accuracy: 0.7890 - loss: 0.4694 - val_accuracy: 0.7672 - val_loss: 0.4976
Epoch 17/100
[1m14/14[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 16ms/step - accuracy: 0.7702 - loss: 0.4673 - val_accuracy: 0.7672 - val_loss: 0.4942
Epoch 18/100
[1m14/14[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 17ms/step - accuracy: 0.7860 - loss: 0.4590 - val_accuracy: 0.7725 - val_loss: 0.4910
Epoch 19/100
[1m14/14[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 19ms/step - accuracy: 0.7816 - loss: 0.4466 - val_accuracy: 0.7725 - val_loss: 0.4874
Epoch 20/100
[1m14/14[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m1s[0m 13ms/step - accuracy: 0.7878 - loss: 0.4385 - val_accuracy: 0.7778 - val_loss: 0.4842
Epoch 21/100
[1m14/14[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 7ms/step - accuracy: 0.7591 - loss: 0.4712 - val_accuracy: 0.7831 - val_loss: 0.4816
Epoch 22/100
[1m14/14[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 7ms/step - accuracy: 0.7941 - loss: 0.4304 - val_accuracy: 0.7831 - val_loss: 0.4786
Epoch 23/100
[1m14/14[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 7ms/step - accuracy: 0.8018 - loss: 0.4240 - val_accuracy: 0.7831 - val_loss: 0.4769
Epoch 24/100
[1m14/14[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 7ms/step - accuracy: 0.7721 - loss: 0.4767 - val_accuracy: 0.7778 - val_loss: 0.4755
Epoch 25/100
[1m14/14[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 7ms/step - accuracy: 0.7649 - loss: 0.4617 - val_accuracy: 0.7672 - val_loss: 0.4736
Epoch 26/100
[1m14/14[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 7ms/step - accuracy: 0.7880 - loss: 0.4148 - val_accuracy: 0.7672 - val_loss: 0.4713
Epoch 27/100
[1m14/14[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 8ms/step - accuracy: 0.7862 - loss: 0.4458 - val_accuracy: 0.7672 - val_loss: 0.4718
Epoch 28/100
[1m14/14[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 7ms/step - accuracy: 0.7971 - loss: 0.4291 - val_accuracy: 0.7672 - val_loss: 0.4703
Epoch 29/100
[1m14/14[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 7ms/step - accuracy: 0.7709 - loss: 0.4286 - val_accuracy: 0.7672 - val_loss: 0.4679
Epoch 30/100
[1m14/14[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 7ms/step - accuracy: 0.7939 - loss: 0.4301 - val_accuracy: 0.7672 - val_loss: 0.4665
Epoch 31/100
[1m14/14[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 7ms/step - accuracy: 0.7871 - loss: 0.4393 - val_accuracy: 0.7672 - val_loss: 0.4665
Epoch 32/100
[1m14/14[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 7ms/step - accuracy: 0.7824 - loss: 0.4321 - val_accuracy: 0.7725 - val_loss: 0.4647
Epoch 33/100
[1m14/14[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 10ms/step - accuracy: 0.7987 - loss: 0.4359 - val_accuracy: 0.7725 - val_loss: 0.4643
Epoch 34/100
[1m14/14[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 7ms/step - accuracy: 0.8198 - loss: 0.4110 - val_accuracy: 0.7725 - val_loss: 0.4635
Epoch 35/100
[1m14/14[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 7ms/step - accuracy: 0.7796 - loss: 0.4268 - val_accuracy: 0.7672 - val_loss: 0.4635
Epoch 36/100
[1m14/14[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 7ms/step - accuracy: 0.7835 - loss: 0.4394 - val_accuracy: 0.7672 - val_loss: 0.4624
Epoch 37/100
[1m14/14[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 10ms/step - accuracy: 0.8138 - loss: 0.4157 - val_accuracy: 0.7619 - val_loss: 0.4616
Epoch 38/100
[1m14/14[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 7ms/step - accuracy: 0.7920 - loss: 0.4308 - val_accuracy: 0.7460 - val_loss: 0.4609
Epoch 39/100
[1m14/14[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 7ms/step - accuracy: 0.7826 - loss: 0.4411 - val_accuracy: 0.7566 - val_loss: 0.4609
Epoch 40/100
[1m14/14[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 7ms/step - accuracy: 0.7653 - loss: 0.4597 - val_accuracy: 0.7513 - val_loss: 0.4598
Epoch 41/100
[1m14/14[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 8ms/step - accuracy: 0.7925 - loss: 0.4191 - val_accuracy: 0.7566 - val_loss: 0.4604
Epoch 42/100
[1m14/14[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 7ms/step - accuracy: 0.8163 - loss: 0.4024 - val_accuracy: 0.7566 - val_loss: 0.4603
Epoch 43/100
[1m14/14[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 7ms/step - accuracy: 0.8015 - loss: 0.4214 - val_accuracy: 0.7513 - val_loss: 0.4591
Epoch 44/100
[1m14/14[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 7ms/step - accuracy: 0.7895 - loss: 0.4372 - val_accuracy: 0.7566 - val_loss: 0.4602
Epoch 45/100
[1m14/14[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 7ms/step - accuracy: 0.8012 - loss: 0.4219 - val_accuracy: 0.7513 - val_loss: 0.4593
Epoch 46/100
[1m14/14[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 10ms/step - accuracy: 0.8030 - loss: 0.4299 - val_accuracy: 0.7513 - val_loss: 0.4584
Epoch 47/100
[1m14/14[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 7ms/step - accuracy: 0.8135 - loss: 0.3938 - val_accuracy: 0.7513 - val_loss: 0.4582
Epoch 48/100
[1m14/14[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 7ms/step - accuracy: 0.8124 - loss: 0.4165 - val_accuracy: 0.7513 - val_loss: 0.4587
Epoch 49/100
[1m14/14[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 7ms/step - accuracy: 0.8037 - loss: 0.4129 - val_accuracy: 0.7513 - val_loss: 0.4581
Epoch 50/100
[1m14/14[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 7ms/step - accuracy: 0.7896 - loss: 0.4202 - val_accuracy: 0.7460 - val_loss: 0.4579
Epoch 51/100
[1m14/14[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 7ms/step - accuracy: 0.7994 - loss: 0.4367 - val_accuracy: 0.7354 - val_loss: 0.4567
Epoch 52/100
[1m14/14[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 7ms/step - accuracy: 0.8106 - loss: 0.4146 - val_accuracy: 0.7513 - val_loss: 0.4584
Epoch 53/100
[1m14/14[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 7ms/step - accuracy: 0.8045 - loss: 0.4196 - val_accuracy: 0.7513 - val_loss: 0.4575
Epoch 54/100
[1m14/14[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 7ms/step - accuracy: 0.8032 - loss: 0.4253 - val_accuracy: 0.7619 - val_loss: 0.4575
Epoch 55/100
[1m14/14[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 7ms/step - accuracy: 0.8206 - loss: 0.3812 - val_accuracy: 0.7354 - val_loss: 0.4562
Epoch 56/100
[1m14/14[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 10ms/step - accuracy: 0.8243 - loss: 0.3859 - val_accuracy: 0.7407 - val_loss: 0.4566
Epoch 57/100
[1m14/14[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 13ms/step - accuracy: 0.8162 - loss: 0.4148 - val_accuracy: 0.7513 - val_loss: 0.4573
Epoch 58/100
[1m14/14[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 10ms/step - accuracy: 0.8217 - loss: 0.4149 - val_accuracy: 0.7407 - val_loss: 0.4566
Epoch 59/100
[1m14/14[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 12ms/step - accuracy: 0.7923 - loss: 0.4262 - val_accuracy: 0.7407 - val_loss: 0.4558
Epoch 60/100
[1m14/14[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 10ms/step - accuracy: 0.7828 - loss: 0.4501 - val_accuracy: 0.7354 - val_loss: 0.4557
Epoch 61/100
[1m14/14[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 10ms/step - accuracy: 0.7971 - loss: 0.4353 - val_accuracy: 0.7407 - val_loss: 0.4550
Epoch 62/100
[1m14/14[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 13ms/step - accuracy: 0.8198 - loss: 0.4084 - val_accuracy: 0.7354 - val_loss: 0.4544
Epoch 63/100
[1m14/14[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 12ms/step - accuracy: 0.8154 - loss: 0.4011 - val_accuracy: 0.7354 - val_loss: 0.4542
Epoch 64/100
[1m14/14[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 9ms/step - accuracy: 0.8137 - loss: 0.4137 - val_accuracy: 0.7354 - val_loss: 0.4547
Epoch 65/100
[1m14/14[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 7ms/step - accuracy: 0.8350 - loss: 0.3822 - val_accuracy: 0.7460 - val_loss: 0.4556
Epoch 66/100
[1m14/14[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 7ms/step - accuracy: 0.7956 - loss: 0.4270 - val_accuracy: 0.7354 - val_loss: 0.4545
Epoch 67/100
[1m14/14[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 7ms/step - accuracy: 0.8196 - loss: 0.4190 - val_accuracy: 0.7407 - val_loss: 0.4538
Epoch 68/100
[1m14/14[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 7ms/step - accuracy: 0.8106 - loss: 0.4313 - val_accuracy: 0.7407 - val_loss: 0.4553
Epoch 69/100
[1m14/14[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 7ms/step - accuracy: 0.8207 - loss: 0.4226 - val_accuracy: 0.7354 - val_loss: 0.4530
Epoch 70/100
[1m14/14[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 7ms/step - accuracy: 0.8334 - loss: 0.3974 - val_accuracy: 0.7407 - val_loss: 0.4534
Epoch 71/100
[1m14/14[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 7ms/step - accuracy: 0.7881 - loss: 0.4449 - val_accuracy: 0.7354 - val_loss: 0.4534
Epoch 72/100
[1m14/14[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 7ms/step - accuracy: 0.8074 - loss: 0.4221 - val_accuracy: 0.7407 - val_loss: 0.4525
Epoch 73/100
[1m14/14[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 7ms/step - accuracy: 0.8267 - loss: 0.4023 - val_accuracy: 0.7460 - val_loss: 0.4528
Epoch 74/100
[1m14/14[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 7ms/step - accuracy: 0.8069 - loss: 0.4141 - val_accuracy: 0.7302 - val_loss: 0.4522
Epoch 75/100
[1m14/14[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 7ms/step - accuracy: 0.8100 - loss: 0.4160 - val_accuracy: 0.7407 - val_loss: 0.4511
Epoch 76/100
[1m14/14[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 7ms/step - accuracy: 0.8153 - loss: 0.4101 - val_accuracy: 0.7302 - val_loss: 0.4511
Epoch 77/100
[1m14/14[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 7ms/step - accuracy: 0.8301 - loss: 0.3827 - val_accuracy: 0.7354 - val_loss: 0.4512
Epoch 78/100
[1m14/14[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 8ms/step - accuracy: 0.8017 - loss: 0.4064 - val_accuracy: 0.7460 - val_loss: 0.4502
Epoch 79/100
[1m14/14[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 7ms/step - accuracy: 0.7842 - loss: 0.4210 - val_accuracy: 0.7354 - val_loss: 0.4504
Epoch 80/100
[1m14/14[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 7ms/step - accuracy: 0.7740 - loss: 0.4259 - val_accuracy: 0.7354 - val_loss: 0.4507
Epoch 81/100
[1m14/14[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 8ms/step - accuracy: 0.8149 - loss: 0.4102 - val_accuracy: 0.7460 - val_loss: 0.4495
Epoch 82/100
[1m14/14[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 7ms/step - accuracy: 0.8165 - loss: 0.3984 - val_accuracy: 0.7407 - val_loss: 0.4501
Epoch 83/100
[1m14/14[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 7ms/step - accuracy: 0.7946 - loss: 0.4188 - val_accuracy: 0.7407 - val_loss: 0.4500
Epoch 84/100
[1m14/14[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 7ms/step - accuracy: 0.8195 - loss: 0.3983 - val_accuracy: 0.7354 - val_loss: 0.4502
Epoch 85/100
[1m14/14[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 7ms/step - accuracy: 0.7944 - loss: 0.4191 - val_accuracy: 0.7407 - val_loss: 0.4495
Epoch 86/100
[1m14/14[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 8ms/step - accuracy: 0.8049 - loss: 0.4199 - val_accuracy: 0.7460 - val_loss: 0.4494
Epoch 87/100
[1m14/14[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 7ms/step - accuracy: 0.7934 - loss: 0.4107 - val_accuracy: 0.7407 - val_loss: 0.4500
Epoch 88/100
[1m14/14[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 7ms/step - accuracy: 0.8140 - loss: 0.3881 - val_accuracy: 0.7407 - val_loss: 0.4494
Epoch 89/100
[1m14/14[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 7ms/step - accuracy: 0.8201 - loss: 0.3780 - val_accuracy: 0.7460 - val_loss: 0.4491
Epoch 90/100
[1m14/14[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 7ms/step - accuracy: 0.8228 - loss: 0.3836 - val_accuracy: 0.7460 - val_loss: 0.4482
Epoch 91/100
[1m14/14[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 7ms/step - accuracy: 0.8314 - loss: 0.3716 - val_accuracy: 0.7407 - val_loss: 0.4505
Epoch 92/100
[1m14/14[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 7ms/step - accuracy: 0.8357 - loss: 0.3735 - val_accuracy: 0.7460 - val_loss: 0.4496
Epoch 93/100
[1m14/14[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 7ms/step - accuracy: 0.8049 - loss: 0.4131 - val_accuracy: 0.7513 - val_loss: 0.4502
Epoch 94/100
[1m14/14[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 8ms/step - accuracy: 0.7860 - loss: 0.4454 - val_accuracy: 0.7460 - val_loss: 0.4500
Epoch 95/100
[1m14/14[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 7ms/step - accuracy: 0.8161 - loss: 0.3831 - val_accuracy: 0.7460 - val_loss: 0.4493
Epoch 96/100
[1m14/14[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 7ms/step - accuracy: 0.8245 - loss: 0.3851 - val_accuracy: 0.7460 - val_loss: 0.4494
Epoch 97/100
[1m14/14[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 7ms/step - accuracy: 0.8321 - loss: 0.3715 - val_accuracy: 0.7460 - val_loss: 0.4503
Epoch 98/100
[1m14/14[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 7ms/step - accuracy: 0.8168 - loss: 0.3952 - val_accuracy: 0.7460 - val_loss: 0.4498
Epoch 99/100
[1m14/14[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 7ms/step - accuracy: 0.7912 - loss: 0.4100 - val_accuracy: 0.7460 - val_loss: 0.4510
Epoch 100/100
[1m14/14[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 7ms/step - accuracy: 0.7942 - loss: 0.4118 - val_accuracy: 0.7513 - val_loss: 0.4502
model.summary()
Model: "sequential"
┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━┓ ┃ Layer (type) ┃ Output Shape ┃ Param # ┃ ┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━┩ │ dense (Dense) │ (None, 10) │ 80 │ ├─────────────────────────────────┼────────────────────────┼───────────────┤ │ dense_1 (Dense) │ (None, 10) │ 110 │ ├─────────────────────────────────┼────────────────────────┼───────────────┤ │ dense_2 (Dense) │ (None, 1) │ 11 │ └─────────────────────────────────┴────────────────────────┴───────────────┘
Total params: 605 (2.37 KB)
Trainable params: 201 (804.00 B)
Non-trainable params: 0 (0.00 B)
Optimizer params: 404 (1.58 KB)
history.history.keys()
dict_keys(['accuracy', 'loss', 'val_accuracy', 'val_loss'])
# Graficar Loss train y Loss test:
plt.plot(history.history['loss'])
plt.plot(history.history['val_loss'])
plt.title('Model loss')
plt.ylabel('Loss')
plt.xlabel('Epoch')
plt.legend(['Train', 'Test'], loc='upper left')
plt.show()
# Probabilidades:
y_prob_train = model.predict(X_train)
y_prob = model.predict(X_test)
# Definición de las clases con umbral:
y_pred_train = np.where(y_prob_train > 0.5, 1, 0)
y_pred = np.where(y_prob > 0.5, 1, 0)
[1m14/14[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 5ms/step
[1m6/6[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 5ms/step
# ------------------------
# Evaluación del modelo
# ------------------------
cm_train = confusion_matrix(y_train, y_pred_train, labels=[0,1])
cm_df_train = pd.DataFrame(cm_train, index=["Real 0", "Real 1"], columns=["Predicho 0", "Predicho 1"])
plt.figure(figsize=(5.2,4.2))
sns.heatmap(cm_train, annot=True, fmt="d", cbar=True, linewidths=.5, cmap="coolwarm")
plt.title("Matriz de confusión - train")
plt.xlabel("Predicho"); plt.ylabel("Real")
plt.tight_layout()
plt.show()
cm = confusion_matrix(y_test, y_pred, labels=[0,1])
cm_df = pd.DataFrame(cm, index=["Real 0", "Real 1"], columns=["Predicho 0", "Predicho 1"])
plt.figure(figsize=(5.2,4.2))
sns.heatmap(cm_df, annot=True, fmt="d", cbar=True, linewidths=.5, cmap="coolwarm")
plt.title("Matriz de confusión - Test")
plt.xlabel("Predicho"); plt.ylabel("Real")
plt.tight_layout()
plt.show()
print("\n=== Reporte de Clasificación - train ===")
print(classification_report(y_train, y_pred_train))
print("\n=== Reporte de Clasificación - test ===")
print(classification_report(y_test, y_pred))
=== Reporte de Clasificación - train ===
precision recall f1-score support
0 0.74 0.92 0.82 201
1 0.91 0.72 0.81 239
accuracy 0.81 440
macro avg 0.82 0.82 0.81 440
weighted avg 0.83 0.81 0.81 440
=== Reporte de Clasificación - test ===
precision recall f1-score support
0 0.69 0.83 0.75 86
1 0.83 0.69 0.75 103
accuracy 0.75 189
macro avg 0.76 0.76 0.75 189
weighted avg 0.76 0.75 0.75 189
Optimización de hiperparámetros:#
from sklearn.metrics import accuracy_score
cantidad_modelos = 10
for i in range(cantidad_modelos):
units = np.random.choice([5, 8, 10, 12, 15, 18, 20, 22, 24], 1)[0]
n_hidden = np.random.choice([1, 2], 1)[0]
activation = np.random.choice(['relu', 'tanh', 'selu', 'elu'], 1)[0]
learning_rate = np.random.choice([0.001, 0.01, 0.1], 1)[0]
batch_size = np.random.choice([16, 32, 64], 1)[0]
optimizer = np.random.choice(['Adam', 'RMSprop'], 1)[0]
print(f'Modelo: {i+1}, Units: {units}, Hidden: {n_hidden}, Activation: {activation}, Learning Rate: {learning_rate}, Optimizer: {optimizer}, Batch Size: {batch_size}')
epochs = 200
# Definir el modelo
best_model = Sequential()
best_model.add(Input(shape=(X.shape[1],)))
# Loop para las capas ocultas:
for _ in range(n_hidden):
best_model.add(Dense(units, activation=activation))
best_model.add(Dropout(0.2))
# Capa de salida:
best_model.add(Dense(1))
# Optimizador:
if optimizer == 'Adam':
optimizer = optimizers.Adam(learning_rate=learning_rate)
else:
optimizer = optimizers.RMSprop(learning_rate=learning_rate)
# Compilar el modelo:
best_model.compile(loss="binary_crossentropy", metrics=["accuracy"], optimizer=optimizer)
# Entrenar el modelo:
history = best_model.fit(X_train, y_train, epochs=epochs,
validation_data=(X_test, y_test),
batch_size=batch_size,
verbose=0)
# Evaluar el modelo con accuracy:
y_prob_train = best_model.predict(X_train)
y_prob = best_model.predict(X_test)
y_pred_train = np.where(y_prob_train > 0.5, 1, 0)
y_pred = np.where(y_prob > 0.5, 1, 0)
accuracy_train = accuracy_score(y_train, y_pred_train.flatten())
accuracy_test = accuracy_score(y_test, y_pred.flatten())
recall_train = recall_score(y_train, y_pred_train.flatten())
recall_test = recall_score(y_test, y_pred.flatten())
precision_train = precision_score(y_train, y_pred_train.flatten())
precision_test = precision_score(y_test, y_pred.flatten())
print(f'Accuracy train: {accuracy_train}, Accuracy test: {accuracy_test}')
print(f'Recall train: {recall_train}, Recall test: {recall_test}')
print(f'Precision train: {precision_train}, Precision test: {precision_test}')
# Graficar Loss train y Loss test:
plt.plot(history.history['loss'])
plt.plot(history.history['val_loss'])
plt.title('Model loss')
plt.ylabel('Loss')
plt.xlabel('Epoch')
plt.legend(['Train', 'Test'], loc='upper left')
plt.show()
# Guardar el modelo:
best_model.save(f"best_model_{i+1}.keras")
Modelo: 1, Units: 15, Hidden: 1, Activation: tanh, Learning Rate: 0.1, Optimizer: Adam, Batch Size: 32
[1m14/14[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 4ms/step
[1m6/6[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 5ms/step
Accuracy train: 0.75, Accuracy test: 0.708994708994709
Recall train: 0.6401673640167364, Recall test: 0.6019417475728155
Precision train: 0.864406779661017, Precision test: 0.8157894736842105
Modelo: 2, Units: 15, Hidden: 2, Activation: tanh, Learning Rate: 0.001, Optimizer: Adam, Batch Size: 16
[1m14/14[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 5ms/step
[1m6/6[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 5ms/step
Accuracy train: 0.7454545454545455, Accuracy test: 0.7142857142857143
Recall train: 0.7907949790794979, Recall test: 0.7864077669902912
Precision train: 0.7529880478087649, Precision test: 0.7168141592920354
Modelo: 3, Units: 12, Hidden: 2, Activation: selu, Learning Rate: 0.1, Optimizer: Adam, Batch Size: 32
[1m14/14[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 5ms/step
[1m6/6[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 5ms/step
Accuracy train: 0.7022727272727273, Accuracy test: 0.656084656084656
Recall train: 0.6694560669456067, Recall test: 0.6504854368932039
Precision train: 0.7547169811320755, Precision test: 0.6979166666666666
Modelo: 4, Units: 12, Hidden: 2, Activation: elu, Learning Rate: 0.001, Optimizer: Adam, Batch Size: 16
[1m14/14[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 5ms/step
[1m6/6[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 5ms/step
Accuracy train: 0.7568181818181818, Accuracy test: 0.7301587301587301
Recall train: 0.7447698744769874, Recall test: 0.7572815533980582
Precision train: 0.7946428571428571, Precision test: 0.75
Modelo: 5, Units: 8, Hidden: 2, Activation: elu, Learning Rate: 0.1, Optimizer: RMSprop, Batch Size: 16
[1m14/14[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 5ms/step
[1m6/6[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 5ms/step
Accuracy train: 0.6931818181818182, Accuracy test: 0.7301587301587301
Recall train: 0.7656903765690377, Recall test: 0.8543689320388349
Precision train: 0.6984732824427481, Precision test: 0.7096774193548387
Modelo: 6, Units: 20, Hidden: 2, Activation: selu, Learning Rate: 0.001, Optimizer: RMSprop, Batch Size: 32
[1m14/14[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 5ms/step
[1m6/6[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 7ms/step
Accuracy train: 0.7477272727272727, Accuracy test: 0.7037037037037037
Recall train: 0.7238493723849372, Recall test: 0.6893203883495146
Precision train: 0.7935779816513762, Precision test: 0.7473684210526316
Modelo: 7, Units: 8, Hidden: 1, Activation: elu, Learning Rate: 0.1, Optimizer: RMSprop, Batch Size: 32
[1m14/14[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 5ms/step
[1m6/6[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 6ms/step
Accuracy train: 0.775, Accuracy test: 0.7248677248677249
Recall train: 0.6569037656903766, Recall test: 0.6407766990291263
Precision train: 0.9022988505747126, Precision test: 0.8148148148148148
Modelo: 8, Units: 22, Hidden: 1, Activation: selu, Learning Rate: 0.001, Optimizer: RMSprop, Batch Size: 16
[1m14/14[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 4ms/step
[1m6/6[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 5ms/step
Accuracy train: 0.7522727272727273, Accuracy test: 0.6984126984126984
Recall train: 0.7364016736401674, Recall test: 0.6796116504854369
Precision train: 0.7927927927927928, Precision test: 0.7446808510638298
Modelo: 9, Units: 10, Hidden: 2, Activation: relu, Learning Rate: 0.01, Optimizer: Adam, Batch Size: 64
[1m14/14[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 7ms/step
[1m6/6[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 7ms/step
Accuracy train: 0.8386363636363636, Accuracy test: 0.7566137566137566
Recall train: 0.7196652719665272, Recall test: 0.6699029126213593
Precision train: 0.9772727272727273, Precision test: 0.8518518518518519
Modelo: 10, Units: 22, Hidden: 2, Activation: elu, Learning Rate: 0.01, Optimizer: RMSprop, Batch Size: 16
[1m14/14[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 6ms/step
[1m6/6[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 5ms/step
Accuracy train: 0.8113636363636364, Accuracy test: 0.7619047619047619
Recall train: 0.6610878661087866, Recall test: 0.6213592233009708
Precision train: 0.9875, Precision test: 0.9142857142857143
Mejor modelo:
from keras.models import load_model
model = load_model("best_model_2.keras")
# Probabilidades:
y_prob_train = model.predict(X_train)
y_prob = model.predict(X_test)
# Definición de las clases con umbral:
y_pred_train = np.where(y_prob_train > 0.5, 1, 0)
y_pred = np.where(y_prob > 0.5, 1, 0)
# ------------------------
# Evaluación del modelo
# ------------------------
cm_train = confusion_matrix(y_train, y_pred_train, labels=[0,1])
cm_df_train = pd.DataFrame(cm_train, index=["Real 0", "Real 1"], columns=["Predicho 0", "Predicho 1"])
plt.figure(figsize=(5.2,4.2))
sns.heatmap(cm_train, annot=True, fmt="d", cbar=True, linewidths=.5, cmap="coolwarm")
plt.title("Matriz de confusión - train")
plt.xlabel("Predicho"); plt.ylabel("Real")
plt.tight_layout()
plt.show()
cm = confusion_matrix(y_test, y_pred, labels=[0,1])
cm_df = pd.DataFrame(cm, index=["Real 0", "Real 1"], columns=["Predicho 0", "Predicho 1"])
plt.figure(figsize=(5.2,4.2))
sns.heatmap(cm_df, annot=True, fmt="d", cbar=True, linewidths=.5, cmap="coolwarm")
plt.title("Matriz de confusión - Test")
plt.xlabel("Predicho"); plt.ylabel("Real")
plt.tight_layout()
plt.show()
print("\n=== Reporte de Clasificación - train ===")
print(classification_report(y_train, y_pred_train))
print("\n=== Reporte de Clasificación - test ===")
print(classification_report(y_test, y_pred))
[1m14/14[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 20ms/step
[1m6/6[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 18ms/step
=== Reporte de Clasificación - train ===
precision recall f1-score support
0 0.74 0.69 0.71 201
1 0.75 0.79 0.77 239
accuracy 0.75 440
macro avg 0.74 0.74 0.74 440
weighted avg 0.74 0.75 0.74 440
=== Reporte de Clasificación - test ===
precision recall f1-score support
0 0.71 0.63 0.67 86
1 0.72 0.79 0.75 103
accuracy 0.71 189
macro avg 0.71 0.71 0.71 189
weighted avg 0.71 0.71 0.71 189
# !pip install keras-tuner -q
import keras_tuner
from keras_tuner.tuners import RandomSearch
from keras_tuner import HyperParameters
import keras
def create_model(hp):
units = hp.Int("units", min_value=5, max_value=24, step=1)
n_hidden = hp.Int("n_hidden", min_value=1, max_value=2, step=1)
learning_rate = hp.Float('learning_rate', min_value=1e-4, max_value=1e-2, sampling='log')
activation = hp.Choice("activation", ['relu', 'tanh', 'selu', 'elu'])
batch_size = hp.Choice("batch_size", [16, 32, 64])
optimizer = hp.Choice('optimizer', values=["adam", "rmsprop"])
# Definir el modelo
best_model = Sequential()
best_model.add(Input(shape=(X.shape[1],)))
# Loop para las capas ocultas:
for _ in range(n_hidden):
best_model.add(Dense(units, activation=activation))
best_model.add(Dropout(0.2))
# Capa de salida:
best_model.add(Dense(1))
# Optimizador:
if optimizer == 'Adam':
best_model.compile(loss="binary_crossentropy", metrics=["accuracy"], optimizer=keras.optimizers.Adam(learning_rate=learning_rate))
else:
best_model.compile(loss="binary_crossentropy", metrics=["accuracy"], optimizer=keras.optimizers.RMSprop(learning_rate=learning_rate))
return best_model
# Configurar la búsqueda de hiperparámetros
tuner = RandomSearch(
create_model,
objective='val_accuracy',
max_trials=10,
executions_per_trial=3,
directory='my_dir',
project_name='Optimización_empresas',
overwrite=True
)
Reloading Tuner from my_dir/Optimización_empresas/tuner0.json
tuner.search(X_train, y_train,
validation_data = (X_test, y_test),
epochs = 100,
batch_size=HyperParameters().Int('batch_size', min_value=16, max_value=128, step=16),
verbose = 0)
# Hiperparámetros óptimos:
best_hps = tuner.get_best_hyperparameters()[0]
print(best_hps.get("units"))
print(best_hps.get("n_hidden"))
print(best_hps.get("activation"))
print(best_hps.get("learning_rate"))
print(best_hps.get("optimizer"))
print(best_hps.get("batch_size"))
23
2
tanh
0.005666324070487912
Adam
64
model = create_model(best_hps)
history = model.fit(X_train, y_train,
validation_data = (X_test, y_test),
epochs = 200,
verbose = 0)
# Graficar Loss train y Loss test:
plt.plot(history.history['loss'])
plt.plot(history.history['val_loss'])
plt.title('Model loss')
plt.ylabel('Loss')
plt.xlabel('Epoch')
plt.legend(['Train', 'Test'], loc='upper left')
plt.show()
# Probabilidades:
y_prob_train = model.predict(X_train)
y_prob = model.predict(X_test)
# Definición de las clases con umbral:
y_pred_train = np.where(y_prob_train > 0.5, 1, 0)
y_pred = np.where(y_prob > 0.5, 1, 0)
# ------------------------
# Evaluación del modelo
# ------------------------
cm_train = confusion_matrix(y_train, y_pred_train, labels=[0,1])
cm_df_train = pd.DataFrame(cm_train, index=["Real 0", "Real 1"], columns=["Predicho 0", "Predicho 1"])
plt.figure(figsize=(5.2,4.2))
sns.heatmap(cm_train, annot=True, fmt="d", cbar=True, linewidths=.5, cmap="coolwarm")
plt.title("Matriz de confusión - train")
plt.xlabel("Predicho"); plt.ylabel("Real")
plt.tight_layout()
plt.show()
cm = confusion_matrix(y_test, y_pred, labels=[0,1])
cm_df = pd.DataFrame(cm, index=["Real 0", "Real 1"], columns=["Predicho 0", "Predicho 1"])
plt.figure(figsize=(5.2,4.2))
sns.heatmap(cm_df, annot=True, fmt="d", cbar=True, linewidths=.5, cmap="coolwarm")
plt.title("Matriz de confusión - Test")
plt.xlabel("Predicho"); plt.ylabel("Real")
plt.tight_layout()
plt.show()
print("\n=== Reporte de Clasificación - train ===")
print(classification_report(y_train, y_pred_train))
print("\n=== Reporte de Clasificación - test ===")
print(classification_report(y_test, y_pred))
[1m14/14[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 5ms/step
[1m6/6[0m [32m━━━━━━━━━━━━━━━━━━━━[0m[37m[0m [1m0s[0m 6ms/step
=== Reporte de Clasificación - train ===
precision recall f1-score support
0 0.76 0.86 0.81 201
1 0.87 0.77 0.82 239
accuracy 0.81 440
macro avg 0.81 0.82 0.81 440
weighted avg 0.82 0.81 0.81 440
=== Reporte de Clasificación - test ===
precision recall f1-score support
0 0.72 0.78 0.75 86
1 0.80 0.75 0.77 103
accuracy 0.76 189
macro avg 0.76 0.76 0.76 189
weighted avg 0.76 0.76 0.76 189