[Python] SVM-Training and Test

To apply my test set to SVM I don't want to use train_test_Split function. I want to apply my test set to svm without training to machine. I wrote a simple code like following one. when I apply my sets to svm the accuracy gets worth but when I use train_test_Split function the accuracy is fine. what's wrong with my code? `

#Training dataset
F_HED = np.array(np.abs(ZZ_1))
F_NonHED = np.array(np.abs(ZZ_N1))
Y1 = np.ones(F_HED.shape[0]) #label 1 for HED
Y2 = np.zeros(F_NonHED.shape[0]) #label 0 for Non-HED


#Combining Train sets
X = np.vstack((F_HED, F_NonHED))
Y = np.hstack((Y1, Y2))


#Testing dataset
X_Test1 = np.array(np.abs(ZZ_X))
X_Test2 = np.array(np.abs(ZZ_NX))
Y_Test1 = np.ones(X_Test1.shape[0]) #label 1 Due to being HED
Y_Test2 = np.zeros(X_Test2.shape[0]) #label 0 Due to being HED


#Combining Test Sets
X_Test = np.vstack((X_Test2, X_Test2))
Y_Test = np.hstack((Y_Test2, Y_Test2))

#Normalizing
Scaler = StandardScaler()
X_Train_S = Scaler.fit_transform(X)
X_Test_S = Scaler.transform(X_Test)

#Classifier
clf = SVC(kernel='rbf', C=10)
clf.fit(X_Train_S, Y)

#Predict part
Y_pred = clf.predict(X_Test_S)


TN, FP, FN, TP = confusion_matrix(Y_Test, Y_pred, labels=[0, 1]).ravel()
print("Confusion Matrix:")
print(f"TN: {TN}, FP: {FP}, FN: {FN}, TP: {TP}")`
Confusion Matrix:
TN: 24, FP: 12, FN: 0, TP: 0


This is the output of my code which I expected that at least FP and FN would be less than 5 but, in that output, the worth thing is with TP which is Zero. In fact, it should be near to 22.

Continue reading...