EMPRR: A high-dimensional EM-based piecewise regression algorithm
Research output: Chapter in Book/Report/Conference proceeding › Article in proceedings › Research › peer-review
We propose a novel general piecewise surface regression model that allows for arbitrary functions to be used in each piece, and arbitrary boundary surfaces between pieces. We also give an EM-based algorithm for this model, EMPRR, that scales to high dimensions. We compare EMPRR's performance with those of model trees and functional trees, two regression tree learning methods, on synthetic piecewise data and benchmark data sets. Our results show that EMPRR outperforms the other two methods on the synthetic data sets and performs competitively on the benchmark data sets while generating accurate and compact models.
Original language | English |
---|---|
Title of host publication | Proceedings of the 2004 International Conference on Machine Learning and Applications, ICMLA '04 |
Editors | M. Kantardzic, O. Nasraoui, M. Milanova |
Number of pages | 8 |
Publication date | 2004 |
Pages | 264-271 |
ISBN (Print) | 0780388232, 9780780388239 |
Publication status | Published - 2004 |
Event | 2004 International Conference on Machine Learning and Applications, ICMLA '04 - Louisville, KY, United States Duration: 16 Dec 2004 → 18 Dec 2004 |
Conference
Conference | 2004 International Conference on Machine Learning and Applications, ICMLA '04 |
---|---|
Land | United States |
By | Louisville, KY |
Periode | 16/12/2004 → 18/12/2004 |
Sponsor | IEEE Systems, Man, and Cybernetics Society, ACM SIDKDD, Association for Machine Learning and Applications, ICMLA, University of Louisville, Dep. of Comput. Eng. and Comput. Sci. |
Series | Proceedings of the 2004 International Conference on Machine Learning and Applications, ICMLA '04 |
---|
ID: 305174093