Math::Histo

Fast, memory-safe C histogramming and statistical computing XS library for Perl, wrapping libhisto.

Installation

cpanm Alien::libhisto
cpanm Math::Histo

Features

• 1D & 2D Histograms: Uniform and arbitrary variable-width binning.
• Ingestion Methods: Scalar, arrayref batch (fill_n), and packed binary double buffer fills.
• Analytics: Mean, variance, skewness, kurtosis, quantiles, mode/FWHM, IQR, MAD, trimmed/winsorized mean.
• Statistical Testing: Two-sample $\chi^2$, Kolmogorov-Smirnov, Wasserstein distance, KL divergence, Bhattacharyya distance.
• Non-Linear Curve Fitting: Levenberg-Marquardt optimizer for Gaussian, exponential, polynomial, Breit-Wigner, and power-law models.
• Streaming Quantile Sketches: DDSketch online dynamic logarithmic binning with relative-error guarantees.
• Zero-Loss Serialization: Endian-independent binary wire format and JSON.