Hyperparameters tuning

paraTuning.py

@@ -0,0 +1,39 @@
+from sklearn.model_selection import GridSearchCV
+
+def SGD_tuning(model,X,Y):
+    params = {
+        "loss" : ["hinge", "log", "squared_hinge", "modified_huber"],
+        "alpha" : [0.0001, 0.001, 0.01, 0.1],
+        "penalty" : ["l2", "l1", "none"],
+    }
+
+    clf = GridSearchCV(model, param_grid=params)
+
+    clf.fit(X, Y)
+    print(clf.best_score_)
+    print(clf.best_estimator_)
+
+def PAC_tuning(model,X,Y):
+    params = {
+        "C" : [1,0.1,0.01,0.001,0.00001,0.000001]
+    }
+
+    clf = GridSearchCV(model,param_grid=params)
+
+    clf.fit(X,Y)
+    print(clf.best_score_)
+    print(clf.best_estimator_)
+
+def MLP_tuning(model,X,Y):
+    params = {
+        "activation" : ["logistic","relu","tanh"],
+        "alpha" : [0.001,0.0001,0.00001],
+        "solver" : ["adam","lbfgs","sgd"],
+        "learning_rate" :["constant","adaptive"]
+    }
+
+    clf = GridSearchCV(model,param_grid=params)
+
+    clf.fit(X,Y)
+    print(clf.best_score_)
+    print(clf.best_estimator_)