DMCA
Linkage and autocorrelation cause feature selection bias in relational learning (2002)
Cached
Download Links
| Venue: | In Proc. of the 19th Intl Conference on Machine Learning |
| Citations: | 119 - 36 self |
BibTeX
@INPROCEEDINGS{Jensen02linkageand,
author = {David Jensen and Jennifer Neville},
title = {Linkage and autocorrelation cause feature selection bias in relational learning},
booktitle = {In Proc. of the 19th Intl Conference on Machine Learning},
year = {2002},
pages = {259--266},
publisher = {Morgan Kaufmann}
}
Share
 |
 |
 |
 |
||
OpenURL
Abstract
Two common characteristics of relational data sets — concentrated linkage and relational autocorrelation — can cause learning algorithms to be strongly biased toward certain features, irrespective of their predictive power. We identify these characteristics, define quantitative measures of their severity, and explain how they produce this bias. We show how linkage and autocorrelation affect a representative algorithm for feature selection by applying the algorithm to synthetic data and to data drawn from the Internet Movie Database. 1.1 Relational Data and Statistical Dependence Figure 1 presents two simple relational data sets. In each
Keyphrases
autocorrelation cause feature selection bias relational learning simple relational data set common characteristic synthetic data representative algorithm predictive power statistical dependence figure data drawn define quantitative measure feature selection relational data set internet movie database relational autocorrelation relational data certain feature
