RFDpy 0.1

A package for machine learning model evaluation and data processing.

RFDpy

RFDpy is a Python package for machine learning model evaluation and data processing. It includes tools for dynamic training processing (DTP) and robust prediction data filtering (RPDF).

Installation

You can install the package using pip:

pip install RFDpy


Usage

from RFDpy import DTP, RPDF
from sklearn.model_selection import train_test_split
from xgboost import XGBRegressor

X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=52)
train_data = list(zip(X_train, y_train))
test_data = list(zip(X_test, y_test))

# Example usage of DTP
def model_factory(regularization):
    return XGBRegressor(n_estimators=100, random_state=52,alpha=regularization)

evaluator = DTP(
    model_factory=model_factory,
    initial_train_data=train_data,
    initial_test_data=test_data,
    x_threshold=0.05,
    initial_lr=0.01,
    lr_decay=0.9,
    initial_regularization=0.001,
    regularization_decay=0.9
)
results = evaluator.run_evaluation_cycles(M=10, th_save_path='path\\to\\save\\cycle_data.txt')


for result in results:
    print(result)

evaluator.export_results_to_txt('path\\to\\save\\evaluation_results.txt', results)
evaluator.export_data_to_txt('path\\to\\save\\detailed_data.txt')

# Example usage of RPDF
rpdf = RPDF(target_occurrences=100, target_probability=0.95, error_margin=0.05)
X_train_updated, y_train_updated = rpdf.update_training_data(X_train, y_train)

xgb_regr = XGBRegressor(n_estimators=100, random_state=52)
xgb_regr.fit(X_train_updated, y_train_updated)
y_pred = xgb_regr.predict(X_test)

satisfactory_indices = rpdf.evaluate_predictions(y_test, y_pred)
satisfactory_values = y_test[satisfactory_indices]

output_file_path = 'path\\to\\save\\output_file.txt'
with open(output_file_path, 'w') as file:
    for value in satisfactory_values:
        file.write(f"{value}\n")

print(f"Filtered true values saved to: {output_file_path}")