bbstat

A lightweight library for Bayesian bootstrapping and statistical evaluation.

Installation

From PyPI:

pip install bbstat

From GitHub source code:

git clone https://github.com/cwehmeyer/bbstat.git
cd bbstat
pip install .

Quickstart

import numpy as np
from bbstat import bootstrap

# Data preparation: simulated income for a small population (e.g., a survey of 25 people)
income = np.array([
    24_000, 26_000, 28_000, 30_000, 32_000,
    35_000, 36_000, 38_000, 40_000, 41_000,
    45_000, 48_000, 50_000, 52_000, 54_000,
    58_000, 60_000, 62_000, 65_000, 68_000,
    70_000, 75_000, 80_000, 90_000, 100_000,
], dtype=np.float64)

# Direct estimate of mean income
print(np.mean(income))  # => 52280.0

# Bootstrapped distribution of the mean income.
distribution = bootstrap(data=income, statistic_fn="mean", seed=1)
print(distribution)  # => BootstrapDistribution(mean=52263.8..., size=1000)

# Summarize the bootstrapped distribution of the mean income.
summary = distribution.summarize(level=0.87)
print(summary)  # => BootstrapSummary(mean=52263.8..., ci_low=46566.8..., ci_high=58453.6..., level=0.87)
print(summary.round())  # => BootstrapSummary(mean=52000.0, ci_low=47000.0, ci_high=58000.0, level=0.87)

API Overview

bootstrap(data, statistic_fn, n_boot=1000, ...)

Performs Bayesian bootstrapping on input data using the given statistic.

• data: 1D NumPy array, or tuple/list thereof
• statistic_fn: string or callable (e.g., "mean", "median", or custom function)
• level: credible interval (default 0.87)
• n_boot: number of bootstrap samples
• seed: random seed (optional)
• blocksize: number of resamples to allocate in one block
• fn_kwargs: optional dictionary with parameters for statistic_fn

Returns a BootstrapResult with:

• .mean: estimated statistic value
• .ci: tuple representing lower and upper bounds of the credible interval
• .level: credible level used
• .n_boot: number of bootstraps performed
• .estimates: array of statistic values computed across the bootstrapped posterior samples

Weighted statistic functions included

The module bbstat.statistics provides a number univariate and bivariate weighted statistics:

• "entropy": bbstat.statistics.compute_weighted_entropy(data, weights)
• "eta_square_dependency": bbstat.statistics.compute_weighted_eta_square_dependency(data, weights)
• "log_odds": bbstat.statistics.compute_weighted_log_odds(data, weights, state: int)
• "mean": bbstat.statistics.compute_weighted_mean(data, weights)
• "median": bbstat.statistics.compute_weighted_median(data, weights)
• "mutual_information": bbstat.statistics.compute_weighted_mutual_information(data, weights)
• "pearson_dependence": bbstat.statistics.compute_weighted_pearson_dependence(data, weights, ddof: int = 0)
• "percentile": bbstat.statistics.compute_weighted_percentile(data, weights, percentile: float)
• "probability": bbstat.statistics.compute_weighted_probability(data, weights, state: int)
• "quantile": bbstat.statistics.compute_weighted_quantile(data, weights, quantile: float)
• "self_information": bbstat.statistics.compute_weighted_self_information(data, weights, state: int)
• "spearman_depedence": bbstat.statistics.compute_weighted_spearman_depedence(data, weights, ddof: int = 0)
• "std": bbstat.statistics.compute_weighted_std(data, weights, ddof: int = 0)
• "sum": bbstat.statistics.compute_weighted_sum(data, weights)
• "variance": bbstat.statistics.compute_weighted_variance(data, weights, ddof: int = 0)

If you want to use your own custom functions, please adhere to this pattern

def custom_statistic(data, weights, *, **kwargs) -> float

where data is

• a 1D numpy array of length n_data or
• a tuple/list of 1D numpy arrays, each of length n_data

and weights is a 1D numpy array of length n_data, with non-negative elements that sum up to one. The function may also take additional parameters which can be supplied via **kwargs.

License

This project is licensed under the MIT License.

Contributing

Contributions are welcome! Please open an issue or submit a pull request.