def svm_skin():
print("\nSVM Classification for Skin Cancer data:\n")
x_train, x_test, y_train, y_test = get_data_skin()
svm = SVM(x_train, np.where(y_train == 0, -1, y_train), 100, 0.01)
y_pred = svm.predict(x_train)
print("\nTraining Classification accuracy: ")
print(100 - 100 * np.sum(np.abs(y_pred - y_train)) / y_pred.shape[0])
confusionMatrix(y_train, y_pred)
y_pred = svm.predict(x_test)
print("\nTesting Classification accuracy: ")
print(100 - 100 * np.sum(np.abs(y_pred - y_test)) / y_pred.shape[0])
confusionMatrix(y_test, y_pred)
print("ROC Curve: ")
plot_roc_curve(y_test, y_pred)
def logistic_skin():
print("\nLogistic Regression for Skin Cancer data:\n")
x_train, x_test, y_train, y_test = get_data_skin()
logistic = Logistic(x_train, y_train)
y_pred = logistic.predict(x_train)
print("\nTraining Classification accuracy: ")
print(100 - 100 * np.sum(np.abs(y_pred - y_train)) / y_pred.shape[0])
confusionMatrix(y_train, y_pred)
y_pred = logistic.predict(x_test)
print("\nTesting Classification accuracy: ")
print(100 - 100 * np.sum(np.abs(y_pred - y_test)) / y_pred.shape[0])
confusionMatrix(y_test, y_pred)
print("ROC Curve: ")
plot_roc_curve(y_test, y_pred)
def predict(self, X_test, y_test):
X_input = autograd.Variable(X_test)
y_target = autograd.Variable(y_test.long())
with torch.no_grad():
y_output = self(X_input)
prediction = y_output.max(1)[1]
FP = (prediction - y_target).nonzero().shape[0]
print(
"\nFeed-Forward NN Test Classification Accuracy: {:.2f}%\n".format(
(1 - FP / (y_target.shape[0])) * 100))
confusionMatrix(y_target, prediction)
print("ROC Curve: ")
plot_roc_curve(y_target, prediction)
def cnn():
x_train, x_test, y_train, y_test = get_data_skin(model='cnn')
x_train = x_train / np.float32(255)
y_train = y_train.astype(np.int32)
x_test = x_test / np.float32(255)
y_test = y_test.astype(np.int32)
# Create the Estimator
skin_cancer_classifier = tf.estimator.Estimator(
model_fn=cnn_model_fn, model_dir="skin_convnet_model")
# Set up logging for predictions
tensors_to_log = {"probabilities": "softmax_tensor"}
logging_hook = tf.train.LoggingTensorHook(tensors=tensors_to_log,
every_n_iter=50)
# Train the model
train_input_fn = tf.estimator.inputs.numpy_input_fn(x={"x": x_train},
y=y_train,
batch_size=100,
num_epochs=None,
shuffle=True)
# train one step and display the probabilties
skin_cancer_classifier.train(input_fn=train_input_fn,
steps=1,
hooks=[logging_hook])
skin_cancer_classifier.train(input_fn=train_input_fn, steps=100)
eval_input_fn = tf.estimator.inputs.numpy_input_fn(x={"x": x_test},
y=y_test,
num_epochs=1,
shuffle=False)
predictions = list(skin_cancer_classifier.predict(input_fn=eval_input_fn))
predict = [i['classes'] for i in predictions]
confusionMatrix(y_test, np.asarray(predict))
print("ROC Curve: ")
plot_roc_curve(y_test, np.asarray(predict))
eval_results = skin_cancer_classifier.evaluate(input_fn=eval_input_fn)
print(eval_results)
def train(self, X_train, y_train):
X_input = autograd.Variable(X_train)
y_target = autograd.Variable(y_train.long())
opt = optim.SGD(params=self.parameters(), lr=0.01)
for epoch in range(max_epochs):
self.zero_grad()
output = self(X_input)
predict = output.max(1)[1]
loss = F.cross_entropy(output, y_target)
loss.backward()
opt.step()
FP = (predict - y_target).nonzero().shape[0]
print("\nFeed-Forward NN Training Classification Accuracy: {:.2f}%".
format((1 - FP / y_target.shape[0]) * 100))
confusionMatrix(y_target, predict)
print("ROC Curve: ")
plot_roc_curve(y_target, predict)