> ## 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.

# Nonparametric estimation

> Kernel density estimates of return distributions (univariate, multivariate and conditional), LOWESS smoothing, and kernel regression (local constant or local linear, optionally censored) with marginal effects.

Toolset `stats_nonparametric`: 4 tools.

| Tool | What it does | Notes |
| - | - | - |
| `stats_kde` | Univariate kernel density estimate of a series (e.g. daily returns): the density curve and its CDF on a grid as a series with charts, the bandwidth used, the mode, and quantiles of the smoothed distribution (e.g. the 1% and 5% tail quantiles). | |
| `stats_lowess` | LOWESS (locally weighted scatterplot smoothing) of y against x, or of a time series against time: the smoothed value for every observation (null where y or x is missing) as a series with a chart, and the residual standard deviation. | |
| `stats_kde_multivariate` | Kernel density of one or more variables, continuous or discrete, estimated jointly f(cols) or conditionally f(cols \| given) (e.g. the density of an asset's return given the market's), with rule-of-thumb or cross-validated bandwidths. | up to 300 s |
| `stats_kernel_regression` | Nonparametric kernel regression of y on x: the conditional mean E\[y \| x] with no functional form assumed, and the marginal effect (partial derivative) of each regressor at every row, local linear or local constant (Nadaraya-Watson, whose effect exists only for a single continuous regressor), with cross-validated or AIC bandwidths. | up to 300 s |

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.