diff --git a/BECK/app.py b/BECK/app.py
index 51e9b7d..e219344 100644
--- a/BECK/app.py
+++ b/BECK/app.py
@@ -52,4 +52,4 @@
     class_comment += 1
     continue
 
-df_cve.to_csv('./datasets/coseno.csv', index=False, encoding="utf-8")
\ No newline at end of file
+df_cve.to_csv('./datasets/coseno.csv', index=False, encoding="utf-8")
diff --git a/BECK/helpers/change_class.py b/BECK/helpers/change_class.py
new file mode 100644
index 0000000..3411d08
--- /dev/null
+++ b/BECK/helpers/change_class.py
@@ -0,0 +1,15 @@
+import pandas as pd
+
+df_completo = pd.read_csv(
+    './comentarios_coseno_distancia.csv', encoding='utf-8', low_memory=False)
+df_comentarios = pd.read_csv(
+    './datasets/comentarios_full.csv', encoding='utf-8', low_memory=False)
+for i in range(0, len(df_completo['Comentario'].index)):
+    # for i in range(0, 3):
+    print(i)
+    clase = df_comentarios[df_comentarios['text'] ==
+                           df_completo.iloc[i]['Comentario']]['class']
+    df_completo.at[i, 'Clase'] = clase.iloc[0]
+
+df_completo.to_csv('./comentarios_coseno_distancia.csv',
+                   encoding='utf-8', index=False)
diff --git a/BECK/helpers/get_item_dataset.py b/BECK/helpers/get_item_dataset.py
index af30c65..6f6d97c 100644
--- a/BECK/helpers/get_item_dataset.py
+++ b/BECK/helpers/get_item_dataset.py
@@ -1,70 +1,42 @@
-from cmath import inf
 import pandas as pd
-from pprint import pprint
-
-df = pd.read_csv('./datasets/coseno_vs_euclidian.csv')
-values_items_largos = [0, 1, 1, 2, 2, 3, 3]
-df_results = pd.read_csv('./results.csv')
+import numpy as np
+df = pd.read_csv('./comentarios_coseno_distancia', encoding='utf-8')
+df_coseno = pd.read_csv(
+    'drive/MyDrive/word_embedding_spanish/dataset_entrenamiento_coseno.csv', encoding='utf-8', delimiter=';')
+array = []
+items = list(df_coseno.columns)
 classes = list(df['Clase'])
-class_idx = 0
-columns = list(df_results.columns)[2:]
-for i in range(0, 7068):
-  column = 0
-  comment = list(df.iloc[i])
-  result = {}
-  result['Comentario'] = comment[0]
-  result['Comentario Preprocesado'] = comment[1]
-  rest_comment = comment[-29:-1]
-  comment = comment[2:-29]
-  for i in range(0, len(comment), 8):
-    data = comment[i:(i+8)]
-    # Euclidian
-    menor_euclidian = inf
-    menor_euclidian_idx = 0
-    #Coseno
-    menor_coseno = inf
-    menor_coseno_idx = 0
-    # Coseno
-    for idx, i2 in enumerate(range(0, len(data), 2)):
-      if data[i2] < menor_coseno:
-        menor_coseno = data[i2]
-        menor_coseno_idx = idx
-    # Euclidian
-    for idx, i3 in enumerate(range(1, len(data), 2)):
-      if data[i3] < menor_euclidian:
-        menor_euclidian = data[i3]
-        menor_euclidian_idx = idx
-    result[columns[column]] = menor_coseno_idx
-    column += 1
-    result[columns[column]] = menor_euclidian_idx
-    column += 1
-  # Anormal items
-  for i in range(0, len(rest_comment), 14):
-    data = rest_comment[i:(i+14)]
-    # Euclidian
-    menor_euclidian = inf
-    menor_euclidian_idx = 0
-    #Coseno
-    menor_coseno = inf
-    menor_coseno_idx = 0
-    # Coseno
-    for idx, i2 in enumerate(range(0, len(data), 2)):
-      if data[i2] < menor_coseno:
-        menor_coseno = data[i2]
-        menor_coseno_idx = idx
-    # Euclidian
-    for idx, i3 in enumerate(range(1, len(data), 2)):
-      if data[i3] < menor_euclidian:
-        menor_euclidian = data[i3]
-        menor_euclidian_idx = idx
-    result[columns[column]] = values_items_largos[menor_coseno_idx]
-    column += 1
-    result[columns[column]] = values_items_largos[menor_euclidian_idx]
-    column += 1
-
-  result['Clase'] = classes[class_idx]
-  class_idx += 1
-  df_results = df_results.append(result, ignore_index=True)
-
-
-df_results.to_csv('results_coseno_euclidian.csv', index=False)
\ No newline at end of file
+items_raros = [0, 1, 1, 2, 2, 3, 3]
+comment_count = 0
+for i in range(0, len(df.index)):
+  comment = list(df.copy().iloc[i])[2:][:-1]
+  normal = list(
+      np.array(comment.copy()[:60] + comment.copy()[67:71] + comment.copy()[78:]))
+  special = list(np.array(comment.copy()[60:67] + comment.copy()[71:78]))
+  comment_data = {}
+  item_count = 0
+  for index in range(0, len(normal), 4):
+    item = normal[index: index + 4]
+    mayor = 0
+    mayor_idx = 0
+    # print(item)
+    for index, result in enumerate(item):
+      if result > mayor:
+        mayor = result
+        mayor_idx = index
+    comment_data[items[item_count]] = mayor_idx
+    item_count += 1
+  for index2 in range(0, len(special), 7):
+    item2 = special[index2: index2 + 7]
+    mayor2 = 0
+    mayor_idx2 = 0
+    for index2, result2 in enumerate(item2):
+      if result2 > mayor2:
+        mayor2 = result2
+        mayor_idx2 = index2
+    comment_data[items[item_count]] = items_raros[mayor_idx2]
+    item_count += 1
+  comment_data['Clase'] = classes[comment_count]
+  comment_count += 1
+  df_coseno = df_coseno.append(comment_data, ignore_index=True)
+df_coseno.to_csv('dataset_entrenamiento.csv', index=False, encoding="utf-8")
diff --git a/BECK/models/SVM/test1.py b/BECK/models/SVM/test1.py
new file mode 100644
index 0000000..219ad47
--- /dev/null
+++ b/BECK/models/SVM/test1.py
@@ -0,0 +1,50 @@
+from sklearn.svm import SVC
+import pandas as pd
+from sklearn.model_selection import train_test_split
+from sklearn.metrics import classification_report, confusion_matrix, f1_score, accuracy_score, precision_score, recall_score
+def train_val_test_split(df, rstate=42, shuffle=True, stratify=None):
+    strat = df[stratify] if stratify else None
+    train_set, test_set = train_test_split(
+        df, test_size=0.25, random_state=rstate, shuffle=shuffle, stratify=strat)
+    strat = test_set[stratify] if stratify else None
+    val_set, test_set = train_test_split(
+        test_set, test_size=0.5, random_state=rstate, shuffle=shuffle, stratify=strat)
+    return (train_set, val_set, test_set)
+
+
+df = pd.read_csv('../../dataset_entrenamiento.csv')
+train_set, val_set, test_set = train_val_test_split(
+    df, stratify='Clase')
+x_train = train_set.drop(labels='Clase', axis=1)
+y_train = train_set['Clase']
+x_test = test_set.drop(labels='Clase', axis=1)
+y_test = test_set['Clase']
+x_val = val_set.drop(labels='Clase', axis=1)
+y_val = val_set['Clase']
+
+
+train_set, val_set, test_set = train_val_test_split(
+    df, stratify='Clase')
+
+svm_clf = SVC(kernel="linear", degree=3, coef0=10, C=20, probability=True)
+svm_clf.fit(x_train, y_train)
+
+#Test
+print(' TEST '.center(50, '#'))
+y_pred = svm_clf.predict(x_test)
+print(classification_report(y_test, y_pred))
+print(confusion_matrix(y_test, y_pred))
+print('F1_Score: ', f1_score(y_test, y_pred))
+print('Precision: ', precision_score(y_test, y_pred))
+print('Accuracy: ', accuracy_score(y_test, y_pred))
+print('Recall: ', recall_score(y_test, y_pred))
+
+#Validation
+print(' VALIDATION '.center(50, '#'))
+y_pred2 = svm_clf.predict(x_val)
+print(classification_report(y_val, y_pred2))
+print(confusion_matrix(y_val, y_pred2))
+print('F1_Score: ', f1_score(y_val, y_pred2))
+print('Precision: ', precision_score(y_val, y_pred2))
+print('Accuracy: ', accuracy_score(y_val, y_pred2))
+print('Recall: ', recall_score(y_val, y_pred2))
\ No newline at end of file