.

evaluation.py

@@ -4,19 +4,32 @@ from tensorflow import keras
 import matplotlib.pyplot as plt
 from keras import backend as K
 from sklearn.metrics import mean_squared_error
-
-# model = keras.models.load_model('suicide_model.h5')
+from tensorflow.keras import layers
+from tensorflow.keras.layers.experimental import preprocessing
+import tensorflow as tf
 
 train = pd.read_csv('train.csv')
 test = pd.read_csv('test.csv')
 validate = pd.read_csv('validate.csv')
 
-print(train)
 # # podział train set
-# X_train = train.loc[:, train.columns != 'suicides_no']
-# y_train = train[['suicides_no']]
-# X_test = test.loc[:, train.columns != 'suicides_no']
-# y_test = test[['suicides_no']]
+X_train = train.loc[:, train.columns != 'suicides_no']
+y_train = train[['suicides_no']]
+X_test = test.loc[:, train.columns != 'suicides_no']
+y_test = test[['suicides_no']]
+
+normalizer = preprocessing.Normalization()
+normalizer.adapt(np.array(X_train))
+
+model = tf.keras.Sequential([
+    normalizer,
+    layers.Dense(units=1)
+])
+
+model.summary()
+
+model.load_weights('suicide_model.h5')
+
 
 # predictions = model.predict(X_test)
 

preprocesing.py

@@ -27,10 +27,10 @@ train, validate, test = np.split(sc.sample(frac=1, random_state=42),
                                  [int(.6*len(sc)), int(.8*len(sc))])
 
 # zapis do plików
-train.to_csv('train.csv')
-validate.to_csv('validate.csv')
-test.to_csv('test.csv')
+train.to_csv('train.csv', index=False)
+validate.to_csv('validate.csv', index=False)
+test.to_csv('test.csv', index=False)
 
-print(train)
-print(validate)
-print(test)
+# print(train)
+# print(validate)
+# print(test)

training.py

@@ -70,6 +70,8 @@ history = model.fit(
     epochs=EPOCHS,
     validation_split=0.2)
 
+model.save_weights('suicide_model.h5')
+
 test_results = {}
 
 test_results['model'] = model.evaluate(
@@ -90,5 +92,4 @@ test_predictions = model.predict(X_test).flatten()
 
 predictions = model.predict(X_test)
 pd.DataFrame(predictions).to_csv('results.csv')
-
-model.save('suicide_model.h5')
+model.summary()