Description:

Linear mappings are omnipresent in data processing analysis ranging from regression to distance metric learning. The interpretation of coefficients from under-determined mappings raises an unexpected challenge when the original modeling goal does not impose regularization. The RegLin package implements a general posterior regularization strategy for inducing unique results.

The benefits are: <1> reflection of data properties such as smoothness in spectrum profiles, <2> easier interpretation of regularized mapping coefficients, and <3> potentially improved mapping quality for unseen test data, i.e. better generalization.

The package also includes a function for standardizing the mapping coefficient vectors by projection to eigenvectors, and an attribute assessment strategy based on sensitivity analysis of the coefficient vectors - these two methods do not require under-determined systems.

The package contains example cases using pinv() and an external linear model (correlative matrix mapping CMM @ mloss.org) for colon cancer gene expression data and for a data base containing near-infrared spectral profiles. See Readme.txt contained in the package.

Changes to previous version:

Version 1.1 (May 23, 2012) memory and time optimizations distderivrel.m now supports assessing the relevance of attribute pairs

Version 1.0 (Nov 9, 2011) * Initial Announcement on mloss.org.

BibTeX Entry: Download

Corresponding Paper BibTeX Entry: Download

Supported Operating Systems: Platform Independent

Data Formats: Matlab

Tags: Regularization, Linear Model

Archive: download here

Comments

No one has posted any comments yet. Perhaps you'd like to be the first?

Leave a comment

You must be logged in to post comments.