> ## Documentation Index
> Fetch the complete documentation index at: https://docs.fincept.in/llms.txt
> Use this file to discover all available pages before exploring further.

# Regression

> Linear models: ordinary, weighted and generalized least squares, AR errors, quantile, robust, recursive, rolling and beta regression, with robust covariances and predictions.

Toolset `stats_regression`: 9 tools.

| Tool | What it does | Notes |
| - | - | - |
| `stats_ols` | Ordinary least squares regression. | core |
| `stats_wls` | Weighted least squares: each row weighted by the weights column (use 1/variance). | |
| `stats_gls` | Generalized least squares with a known diagonal error covariance (one variance per row). | |
| `stats_glsar` | Feasible GLS with AR(p) errors (iterated Cochrane-Orcutt): regression for time series whose residuals are autocorrelated. | |
| `stats_quantile_regression` | Quantile regression at one or more quantiles (0.5 = median/least absolute deviation): how the regressors move each part of the response's distribution. | |
| `stats_robust_regression` | Robust linear regression by M-estimation (iteratively reweighted least squares): coefficients that outliers cannot drag. | |
| `stats_recursive_ls` | Recursive least squares: the coefficients re-estimated as each observation is added, with the CUSUM and CUSUM-of-squares statistics for parameter stability. | |
| `stats_rolling_regression` | Rolling-window least squares: coefficients (e.g. a stock's beta) and R² re-estimated over a moving window. | |
| `stats_beta_regression` | Beta regression for a response strictly between 0 and 1 (rates, shares, recovery rates): a logit-link mean model with a precision parameter. | |

Inputs, limits and outputs are described in [Fincept Stats](/guides/fincept-stats). Full schemas: `fincept_describe_tool`.


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.