Autocorrelation and linkage cause bias in evaluation of relational learners (2002)
Cached
Download Links
- [kdl.cs.umass.edu]
- [kdl.cs.umass.edu]
- [www.cs.purdue.edu]
- [kdl.cs.umass.edu]
- DBLP
Other Repositories/Bibliography
| Venue: | In Proceedings of the Twelfth International Conference on Inductive Logic Programming |
| Citations: | 9 - 3 self |
BibTeX
@INPROCEEDINGS{Jensen02autocorrelationand,
author = {David Jensen and Jennifer Neville},
title = {Autocorrelation and linkage cause bias in evaluation of relational learners},
booktitle = {In Proceedings of the Twelfth International Conference on Inductive Logic Programming},
year = {2002},
pages = {101--116},
publisher = {Springer}
}
Bookmark
 |
 |
 |
 |
 |
 |
OpenURL
Abstract
Two common characteristics of relational data sets — concentrated linkage and relational auto-correlation — can cause traditional methods of evaluation to greatly overestimate the accuracy of induced models on test sets. We identify these characteristics, define quantitative measures of their severity, and explain how they produce this bias. We show how linkage and autocorrelation affect estimates of model accuracy by applying FOIL to synthetic data and to data drawn from the Internet Movie Database. We show how a modified sampling procedure can eliminate the bias.
