From this page you can download the MATLAB function tvc, which implements the TVC regression model proposed by Ciapanna and Taboga (2011 - "Bayesian Analysis of Coefficient Instability in Dynamic Regressions" - download the paper).
The regression model features time-varying regression coefficients. It is Bayesian, but the specification of priors is uninformative and fully automatic, in the sense that the final user is not required to input any prior. The priors are derived automatically from elementary requirements such as scale-invariance (results must not change when you multiply or divide one or more regressors by a constant).
The algorithm is fast (few milliseconds for moderately sized datasets) and its performance has been thoroughly tested in Monte Carlo experiments.
Download the ZIP file containing the MATLAB routines and start from the routine tvc_demo. Note that this distribution is still quite preliminary.
The research topic of the paper is best described by the following keywords:
Salient features of the model are:
Most of the learning materials found on this website are now available in a traditional textbook format.
