zerowing-ex / Machine_Learning

The individual implementations of various algorithms in ML without any external libraries except NumPy. You can use my code without any limits.

Machine_Learning Repository Info

The individual implementations of various algorithms in ML without any external libraries except NumPy.

You can use my code without any limits.

Implemented Algorithms

Logistic

Bayes Nework

Naive Bayes

Neural Network

Algorithm design refers to Keras and TensorFlow. The construction way of the NN only supports Function API.

nn_input = Input(shape=(x_train.shape[1]))
x = Dense(1024, activation='relu')(nn_input)
x = BatchNormalization()(x)
x = Dense(512, activation='relu')(x)
x = BatchNormalization()(x)
nn_output = Dense(10, activation='softmax')(x)

Activations

• Activation

Abstract class, which is the base class of all activation classes

• Including Indentity, Softmax, Sigmoid, tanh, ReLU, LeakyReLU

Initializers

• Initializer

Abstract class, which is the base class of all initializer classes

• Including Zeros, Ones, RandomUniform, RandomNormal

Layers

Base Layers

• Layer

Abstract class, which is the base class of all layers

• Input

The input of the nerual network

• Dense

Namely, fully connected layer

Normalization Layers

• BatchNormalization

Regularization Layers

• Dropout (Not implemented)

Losses

• LossFunction

Abstract class, which is the base class of all loss functions

Probabilistic Losses

• CategoricalCrossentropy
• SparseCategoricalCrossentropy (Not implemented)

Regression Losses

• MeanSquaredError (Not implemented)

Metrics

• Metric

Abstract class, which is the base class of all metrics

• Accuracy

Optimizers

• Optimizer

Abstract class, which is the base class of all optimizers

• SGD

Stochastic Gradient Descent, momentum and Nestrov is not supported yet

Model

• Model

compile -> fit -> score/predict (not implemented)

The parameter 'batch_size' is kind of meaningless of CPU training but just for fun :D.

model = Model(inputs=nn_input, outputs=nn_output)
model.compile(optimizer='sgd', loss='categorical_crossentropy', metrics='acc')
model.fit(x_train, y_train, validation_data=(x_valid, y_valid), batch_size=128)

Utils

• to_one_hot function

Convert numpy.ndarray type to the one-hot form

Support Vector Machine (Not implemented)