Crate inferust 

inferust

Statistical modeling for Rust — a statsmodels-inspired library.

Modules

Module	Contents
regression	OLS/WLS/GLS/FGLS, quantile regression, plus rolling and recursive OLS with fast/stable/HAC solvers, HC0-HC3 and Newey-West SEs, confidence intervals, influence diagnostics, full summary
glm	Binary logistic and Poisson regression with Wald inference, LRT, residual diagnostics, prediction intervals, classification metrics
gam	Gaussian additive models with spline basis expansion
gmm	Instrumental variables / 2SLS econometrics starter
survival	Kaplan-Meier estimator, log-rank test, Cox proportional hazards regression
statespace	Scalar Kalman filter and local-level state-space model
time_series	ARIMA/SARIMA/SARIMAX via CSS, VAR/VECM/VARMAX, AR, ACF/PACF, Ljung-Box, ADF unit root, KPSS stationarity
hypothesis	t-tests, chi-squared, ANOVA, Mann-Whitney U, Kruskal-Wallis, KS tests, Shapiro-Wilk
contingency	2x2 tables, odds/risk ratios, McNemar, CMH
diagnostics	VIF, Breusch-Pagan, White, and RESET diagnostics
discrete	Probit, ordered logit, negative binomial, multinomial logit, and zero-inflated Poisson
glm_family	Generic Gaussian/Binomial/Poisson GLM front-end
multivariate	MANOVA and PCA starters
imputation	Mean imputation and MICE-style chained equations
treatment	Propensity scores, IPW treatment effects, and balance diagnostics
evaluation	Regression/classification metrics and bootstrap intervals
graphics	Dependency-light SVG plots for lines, scatter, residuals, and ACF bars
plot	Full-featured SVG/ASCII plot builder: line, scatter, bar, step, band, ACF, survival, and residual charts
robust	Huber robust linear regression
gee	Independence-working-correlation GEE
mixed	Random-intercept mixed linear model
descriptive	Summary stats (mean, std, skewness, kurtosis, quartiles)
data	Named-column DataFrame with formula API: y ~ C(g) + x1*x2 - 1 + offset(e)
correlation	Pearson, Spearman, correlation matrices

OLS covariance options

regression::Ols defaults to classical (homoskedastic) standard errors. Switch with .robust() (HC1), .with_covariance(OlsCovariance::Hc3), or .hac(lags) (Newey-West) for time series regressions.

Formula syntax

data::DataFrame accepts R-style formulas:

• "y ~ x1 + x2" — main effects
• "y ~ x1 + x2 - 1" — no intercept
• "y ~ C(group) + x1" — inline one-hot encoding
• "y ~ x1:x2" — interaction term
• "y ~ x1 * x2" — main effects + interaction
• "y ~ log(x) + sqrt(z)" — numeric transforms
• "y ~ x + offset(exp)" — Poisson offset

For Rust-native ergonomics, the formula! macro turns formula-like tokens into a formula string:

let f = inferust::formula!(y ~ x1 + C(group));
assert_eq!(f, "y ~ x1 + C(group)");

Quick start

use inferust::regression::Ols;

let x = vec![vec![1.0], vec![2.0], vec![3.0], vec![4.0], vec![5.0]];
let y = vec![2.1, 3.9, 6.2, 7.8, 10.1];

Ols::new()
    .with_feature_names(vec!["hours".to_string()])
    .fit(&x, &y)
    .unwrap()
    .print_summary();

Re-exports

pub use error::InferustError;

pub use error::Result;

Modules

contingency

Contingency table statistics.

correlation

data

descriptive

diagnostics

discrete

error

evaluation

gam

Generalized Additive Model starters.

gee

glm

glm_family

gmm

Generalized Method of Moments and instrumental-variable estimators.

graphics

Lightweight plotting helpers that render statistical graphics as SVG.

hypothesis

imputation

Multiple-imputation starters.

mixed

multivariate

Multivariate statistics.

plot

Plotting utilities — SVG file output and ASCII terminal charts.

regression

robust

statespace

State-space models and Kalman filtering.

survival

Survival analysis: Kaplan-Meier estimator and Cox proportional hazards regression.

time_series

Time series modelling and diagnostics.

treatment

Treatment-effect estimators and balance diagnostics.

Macros

formula

Build a formula string from Rust tokens.