Add adam and ridge classifier

tests/models/logistic_regression.py

@@ -1,10 +1,25 @@
 from sklearn.linear_model import LogisticRegression
+from sklearn.neural_network import MLPClassifier
+
 
 def logistic_regression(X_train, y_train, X_test):
   print("[Executing]: Running Logistic Regression model ...\n")
 
   # Model Fitting
-  model = LogisticRegression(random_state = 0, solver='lbfgs', multi_class='multinomial')
+  model = LogisticRegression(random_state = 21)
+  model.fit(X_train, y_train)
+
+  # Predictions on test dataset
+  y_pred = model.predict(X_test)
+
+  return y_pred
+
+
+def logistic_regression_adam(X_train, y_train, X_test):
+  print("[Executing]: Running Logistic Regression model with Adam ...\n")
+
+  # Model Fitting
+  model = MLPClassifier(hidden_layer_sizes=(1), solver='adam', random_state=48)
   model.fit(X_train, y_train)
 
   # Predictions on test dataset

tests/models/parent_model.py

@@ -32,12 +32,12 @@ def execute_model(X, y, num_splits, seed, model, with_undersampling = False, maj
       print("[Testing]: Count of test data before Undersampling = ", X_train.shape[0])
       verdict = undersampling_method.undersample(X_train, y_train, majority_class)
 
-      # X_train = X_train[verdict, :]
-      # y_train = y_train[verdict]
+      X_train = X_train[verdict, :]
+      y_train = y_train[verdict]
 
       # In-buit near miss algorithm
-      nr = NearMiss()
-      X_train, y_train = nr.fit_resample(X_train, y_train)
+      # nr = NearMiss()
+      # X_train, y_train = nr.fit_resample(X_train, y_train)
 
       # Note: Be careful while plotting, make sure same features are being compared
       # plt.scatter(X_train[:, 0], X_train[:, 1], marker = '.', c = y_train)
@@ -66,7 +66,7 @@ def execute_model(X, y, num_splits, seed, model, with_undersampling = False, maj
   class1_metrics_list = np.mean(class1_metrics_list, axis = 0)
 
 
-  print("Majority Class = Class ", majority_class, "\n")
+  print("\nMajority Class = Class ", majority_class)
   print("\n---------------  Cross-validated Evaluation Metrics  ---------------\n")
   print("Accuracy \t= \t", metrics_list[0])
   print("Precision \t= \t", metrics_list[1])

tests/models/ridge_classification.py

@@ -0,0 +1,13 @@
+from sklearn.linear_model import RidgeClassifier
+
+def ridge_classification(X_train, y_train, X_test):
+  print("[Executing]: Running Ridge Regression model ...\n")
+
+  # Model Fitting
+  model = RidgeClassifier(random_state=88)
+  model.fit(X_train, y_train)
+
+  # Predictions on test dataset
+  y_pred = model.predict(X_test)
+
+  return y_pred

tests/sample_test_1.py

@@ -7,7 +7,8 @@
 # from cobraclassifier import classifier_cobra
 # from cobraclassifier import near_miss_v1, near_miss_v2, near_miss_v3, knn_und, edited_knn, condensed_knn, tomek_link
 
-from models.logistic_regression import logistic_regression
+from models.logistic_regression import logistic_regression, logistic_regression_adam
+from models.ridge_classification import ridge_classification
 from models.adaboost import adaboost_classifier
 from models.parent_model import execute_model
 from undersampling_algorithms import near_miss_v1, near_miss_v2, near_miss_v3, knn_und, edited_knn, condensed_knn, tomek_link
@@ -257,9 +258,11 @@ def main():
   X, y, majority_class_label = prepare_data(seed, choice=ch)
 
   # models = [logistic_regression, adaboost_classifier, classifier_cobra.execute_cobra]
-  models = [classifier_cobra.execute_cobra]
+  # models = [classifier_cobra.execute_cobra]
   # models = [logistic_regression, adaboost_classifier]
   # models = [logistic_regression]
+  # models = [logistic_regression_adam]
+  models = [ridge_classification]
 
   for m in models:
     print("\n\n#############################  MODEL -", m.__name__, "  #############################")