extrastats 0.1.0

A Python package for advanced statistical analysis.

extrastats

Advanced statistical tools and routines for Python

extrastats is a Python library that provides high-quality statistical methods to address gaps in mainstream libraries like NumPy, SciPy, and statsmodels. Designed for data scientists, statisticians, and researchers, extrastats includes robust, customizable, and performance-optimized routines.

Copyright 2022-2025 Jerrad Michael Genson

This library is licensed under the Mozilla Public License, v. 2.0.

---

Features

Robust Outlier Detection

• Adjusted Boxplot: A method that extends traditional boxplots using the medcouple statistic for skewness adjustment.

Advanced Permutation Testing

• Flexible resampling strategies:
  • Pairings, samples, independent shuffles, and bootstrapping.
• Parallel computation with joblib.

Confidence Interval Estimation

• Bootstrap-based confidence intervals for arbitrary statistics.
• Support for multiple confidence levels in a single call.

Sample Size Estimation

• Monte Carlo simulations to determine required sample sizes for target confidence interval widths.

Tail Weight Analysis

• Evaluate the tail weight of distributions using the L/RMC method.

Mutual Information

• Compute mutual information for discrete variables.
• Optional normalization for interpretability.

Additional Metrics

• Geometric Coefficient of Variation (GCV)
• Harmonic Variability (HVAR)
• Trimmed statistics and probability operations.

---