Distribution Fitting

v0.4.0

Maximum-likelihood fitting of candidate scipy distributions with AIC/BIC ranking, quantile-quantile data and Kolmogorov-Smirnov goodness-of-fit.

Overview

Module-level functions fit_distributions, qq_data and goodness_of_fit answer the two classic questions of univariate modelling: what distribution does my data follow? and is this specific distribution a good fit? All fitting uses scipy's MLE estimators — no new dependencies.

fit_distributions

fit_distributions(data, column=None, candidates=None, criterion='aic') -> DistributionFitResult

DistributionFitResult

qq_data

qq_data(data, column=None, dist='norm', params=None, n_points=100) -> DataFrame

DataFrame (attrs: r2, distribution)

goodness_of_fit

goodness_of_fit(data, column=None, dist='norm', params=None, alpha=0.05) -> TestResult

TestResult

DistributionFitResult

best_name · best_params · best_dist · criterion · table · to_dataframe() · to_dict() · to_markdown() · to_json()