This section collects various methods in nonparametric statistics. This includes currently kernel density estimation for univariate data, and lowess.
sandbox.nonparametric contains additional functions that are work in progress or don’t have unit tests yet. We are planning to include here nonparametric density estimators, especially based on kernel or orthogonal polynomials, smoothers, and tools for nonparametric models and methods in other parts of statsmodels.
Currently, the public functions and classes are
smoothers_lowess.lowess(endog, exog[, frac, it]) | LOWESS (Locally Weighted Scatterplot Smoothing) |
kde.KDE(endog) | Kernel Density Estimator |
helper functions for kernel bandwidths
bandwidths.bw_scott(x) | Scott’s Rule of Thumb |
bandwidths.bw_silverman(x) | Silverman’s Rule of Thumb |
bandwidths.select_bandwidth(x, bw, kernel) | Selects bandwidth for a selection rule bw |