mlprimer 0.1.0

Reusable machine learning toolkit for summary statistics, correlation, preprocessing, and model selection.

mlprimer

mlprimer is a Python package that assists with early-stage machine learning workflows. It helps users perform data profiling, feature correlation testing, preprocessing with class balancing, model training, and evaluation across common classifiers.

Features

• Generate summary statistics from cleaned datasets
• Identify meaningful feature-target relationships
• Apply SMOTE to handle class imbalance
• Compare baseline classification models
• Perform hyperparameter tuning with GridSearch and RandomizedSearch

Installation

📦 From GitHub (direct)

pip install git+https://github.com/Kingkong2408/mlprimer.git

Usage

from mlprimer import (
    calculate_summary,
    check_correlation,
    apply_smote,
    split_data,
    train_models,
    evaluate_models
)

# Summary
summary = calculate_summary(df)

# Correlation
correlation = check_correlation(df, target="target_column")

# Preprocessing
X_res, y_res = apply_smote(X, y)
X_train, X_test, y_train, y_test = split_data(X_res, y_res)

# Modeling
models = train_models(X_train, y_train)
results = evaluate_models(models, X_test, y_test)
print(results)

Development Goals

• Add support for regression models
• Include more preprocessing tools (e.g., imputation, encoding helpers)
• Improve pipeline integration and configuration support

License

MIT