statsmodels nonparametric estimation

by Georgi Panterov for Python Software Foundation

Statsmodels is a pure Python-based statistics and econometrics package that has drawn significant attention from applied practitioners from the fields of Finance, Economics and social sciences. Many of the basic econometric methods have been developed. In addition some impressive work has been done in developing the time series methods, VARs, and DSGE models to name a few. This GSoC I intend to develop the nonparametric capabilities of statsmodels by focusing in particular on data-driven bandwidth selection procedures, conditional and unconditional multivariate probability and cumulative density estimation and implementing popular nonparametric and semiparametric regression models.