Specificities = []
kf = KFold(n_splits=10)
for train_i, test_i in kf.split(X):
x_train = X[train_i].reshape(X[train_i].shape[0],
X[train_i].shape[1] * X[train_i].shape[2])
x_test = X[test_i].reshape(X[test_i].shape[0],
X[test_i].shape[1] * X[test_i].shape[2])
y_train, y_test = Y[train_i], Y[test_i]
svm_model = SVC(C=100, kernel='linear', gamma='scale')
svm_model.fit(x_train, y_train)
y_pred = svm_model.predict(x_test)
print("-------------------------------")
ACC, TPR, TNR, PPV, NPV, FPR = per.GetPerformanceMetrics(y_test,
y_pred,
weighted=True)
Accuracies.append(ACC)
Recalls.append(TPR)
Specificities.append(TNR)
Precisions.append(PPV)
NPVs.append(NPV)
FPRs.append(FPR)
print("Accuracy: ", ACC)
print("Recall: ", TPR)
print("Specificity: ", TNR)
print("Precision: ", PPV)
print("Negative Predictive Value: ", NPV)
print("FP rate(fall-out): ", FPR)
print(confusion_matrix(y_test, y_pred))
])
model.compile(optimizer="adam",
loss="sparse_categorical_crossentropy",
metrics=['accuracy'])
model.summary()
model.fit(train_x, train_y, verbose=2, batch_size=2, epochs=50)
model.evaluate(test_x, test_y, batch_size=2, verbose=2)
ypred = np.argmax(model.predict(test_x, batch_size=2), axis=-1)
print("-------------------------------")
ACC, TPR, TNR, PPV, NPV, FPR = per.GetPerformanceMetrics(test_y,
ypred,
weighted=True)
print("Accuracy: ", ACC)
print("Recall: ", TPR)
print("Specificity: ", TNR)
print("Precision: ", PPV)
print("Negative Predictive Value: ", NPV)
print("FP rate(fall-out): ", FPR)
print(confusion_matrix(test_y, ypred))
print("-------------------------------")
print(
"0: fighting\n 1: front\n 2: ready\n 3: cat\n 4: horse \n 5: hicho \n 6: seiza"
)