Welcome to the InfoSci Platform
On Bias-Variance Analysis for Probabilistic Logic Models

On Bias-Variance Analysis for Probabilistic Logic Models

Huma Lodhi
Copyright: © 2008 |Volume: 1 |Issue: 3 |Pages: 14
ISSN: 1938-7857|EISSN: 1938-7865|ISSN: 1938-7857|EISBN13: 9781615205462|EISSN: 1938-7865|DOI: 10.4018/jitr.2008070103
Cite Article Cite Article

MLA

Lodhi, Huma. "On Bias-Variance Analysis for Probabilistic Logic Models." JITR vol.1, no.3 2008: pp.27-40. http://doi.org/10.4018/jitr.2008070103

APA

Lodhi, H. (2008). On Bias-Variance Analysis for Probabilistic Logic Models. Journal of Information Technology Research (JITR), 1(3), 27-40. http://doi.org/10.4018/jitr.2008070103

Chicago

Lodhi, Huma. "On Bias-Variance Analysis for Probabilistic Logic Models," Journal of Information Technology Research (JITR) 1, no.3: 27-40. http://doi.org/10.4018/jitr.2008070103

Export Reference

Mendeley
Favorite Full-Issue Download

Abstract

The article introduces bias-variance decomposition in probabilistic logic learning. We use Stochastic Logic Programs for probabilistic logic representation. In order to learn probabilistic logic models we use Failure Adjusted Maximization (FAM) that is an instance of the Expectation Maximization (EM) algorithm for first order logic. Experiments are carried out by concentrating on one kind of application: quantitative modelling of metabolic pathways that is a complex and challenging task in computational systems biology. We apply bias-variance definitions to analyze quantitative modelling of amino acid pathways of Saccharomyces cerevisiae (yeast). The results show the phenomenon of bias-variance trade-off in probabilistic logic learning.

Request Access

You do not own this content. Please login to recommend this title to your institution's librarian or purchase it from the IGI Global bookstore.