Abstract
Selecting an optimal subset of k out of d features for linear regression models given n training instances is often considered intractable for feature spaces with hundreds or thousands of dimensions. We propose an efficient massively-parallel implementation for selecting such optimal feature subsets in a brute-force fashion for small k. By exploiting the enormous compute power provided by modern parallel devices such as graphics processing units, it can deal with thousands of input dimensions even using standard commodity hardware only. We evaluate the practical runtime using artificial datasets and sketch the applicability of our framework in the context of astronomy.
Original language | English |
---|---|
Title of host publication | 2017 IEEE Symposium Series on Computational Intelligence (SSCI) Proceedings |
Number of pages | 8 |
Publisher | IEEE |
Publication date | 2017 |
Pages | 1-8 |
ISBN (Electronic) | 978-1-5386-2726-6 |
DOIs | |
Publication status | Published - 2017 |
Event | 2017 IEEE Symposium Series on Computational Intelligence (SSCI) - Honolulu, United States Duration: 27 Nov 2017 → 1 Dec 2017 |
Conference
Conference | 2017 IEEE Symposium Series on Computational Intelligence (SSCI) |
---|---|
Country/Territory | United States |
City | Honolulu |
Period | 27/11/2017 → 01/12/2017 |
Keywords
- graphics processing units
- least squares approximations
- optimisation
- parallel processing
- regression analysis
- sensitivity analysis
- input dimensions
- linear regression models
- massively-parallel best subset selection
- optimal feature subsets
- optimal subset
- ordinary least-squares regression
- subset selection
- Computational modeling
- Graphics processing units
- Instruction sets
- Optimization
- Runtime
- Task analysis
- Training