hal.ml package

Submodules

hal.ml.correlation module

Correlate values in arrays producing fancy good-looking matrices

class hal.ml.correlation.CorrelationMatrix(title, headers_to_test, headers, data)

Bases: object

Common operations for a correlation matrix

static get_correlation_matrix(matrix)

Finds correlation matrix of matrix

Parameters:matrix – List of features to get correlation matrix Returns:correlation matrix

get_correlation_matrix_from_columns()

Computes correlation matrix of columns :return: Correlation matrix of columns

static save_correlation_matrix_from_folder(folder_path)

Saves each file’s correlation matrix of common headers

Parameters:folder_path – Folder containing logs data

save_to_file(out_file)

Saves correlation matrix of selected headers

Parameters:out_file – Output file

show_correlation_matrix(correlation_matrix)

Shows the given correlation matrix as image

Parameters:correlation_matrix – Correlation matrix of features

show_correlation_matrix_from_columns()

Shows the correlation matrix of columns

hal.ml.features module

Collection of methods to find weights of features and select the best ones

class hal.ml.features.FeatureSelect(x, y)

Bases: object

Selects best features

get_best()

Finds the optimal number of features :return: optimal number of features and ranking

select_k_best(k)

Selects k best features in dataset

Parameters:k – features to select Returns:k best features

hal.ml.predict module

General model to make prediction about everything

class hal.ml.predict.BasePrediction(model, rounds)

Bases: object

The mother of all predictions

train(x_data, y_data)

Trains model on inputs

Parameters:

• x_data – x matrix
• y_data – y array