Search Results for "method"

Orthogonal method for the Identification of Volterra series. It is an extension of Lee-Schetzen method with two major improvments: 1. Reduced identification uncertainty in diagonal kernel points. 2. Possibility to identify each Volterra kernel with an input with different variance. This feature reduces the identification noise on lower order kernels and improve the "resolution" on higher order kernels.

A MATLAB spectral clustering package to handle large data sets (200,000 RCV1 data) on a 4GB memory general machine. We implement various ways of approximating the dense similarity matrix, including nearest neighbors and the Nystrom method.