polars-statistics API Reference¶
Complete API documentation for polars-statistics.
Quick Links¶
| Category | Description |
|---|---|
| Conventions | API patterns, column references, return types |
| Statistical Tests | |
| Parametric Tests | t-tests, Brown-Forsythe, Yuen |
| Non-Parametric Tests | Mann-Whitney, Wilcoxon, Kruskal-Wallis, Brunner-Munzel |
| Distributional Tests | Shapiro-Wilk, D'Agostino normality tests |
| Forecast Tests | Diebold-Mariano, Clark-West, SPA, MCS |
| Correlation Tests | Pearson, Spearman, Kendall, distance correlation, partial correlation |
| Categorical Tests | Chi-square, Fisher exact, McNemar, Cohen's Kappa |
| TOST Equivalence | Two One-Sided Tests for equivalence |
| Regression | |
| Linear Models | OLS, Ridge, Elastic Net, WLS, RLS, BLS, NNLS, Quantile, Isotonic |
| GLM Models | Logistic, Poisson, Negative Binomial, Tweedie, Probit, Cloglog |
| ALM | Augmented Linear Model (24+ distributions) |
| Dynamic Models | LmDynamic time-varying coefficients |
| Demand Classification | AID demand patterns and anomaly detection |
| Formula Syntax | R-style formulas with interactions and polynomials |
| Summary & Predict | Coefficient tables and prediction intervals |
| Diagnostics | Condition number, quasi-separation detection |
| Model Classes | |
| Linear Model Classes | OLS, Ridge, ElasticNet, WLS, RLS, BLS, Quantile, Isotonic |
| GLM Model Classes | Logistic, Poisson, NegativeBinomial, Tweedie, Probit, Cloglog |
| ALM Class | Augmented Linear Model class |
| LmDynamic Class | Dynamic linear model class |
| Aid Class | Demand classification class |
| Test Classes | Statistical test classes |
| Bootstrap Classes | Stationary and circular block bootstrap |
| Reference | |
| Output Structures | All return type definitions |
Installation¶
Basic Usage¶
All functions work as Polars expressions:
import polars as pl
import polars_statistics as ps
df = pl.DataFrame({
"group": ["A"] * 50 + ["B"] * 50,
"y": [...],
"x1": [...],
})
# Run regression per group
result = df.group_by("group").agg(
ps.ols("y", "x1").alias("model")
)
# Extract results
result.with_columns(
pl.col("model").struct.field("r_squared"),
)
See Also¶
- README - Quick start guide
- Polars Documentation