Kernel Matrix-Based Heuristic Multiple Kernel Learning
Kernel theory is a demonstrated tool that has made its way into nearly all areas of machine learning. However, a serious limitation of kernel methods is knowing which kernel is needed in practice. <i>Multiple kernel learning</i> (MKL) is an attempt to learn a new tailored kernel through...
| Main Authors: | , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2022-06-01
|
| Series: | Mathematics |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2227-7390/10/12/2026 |
| _version_ | 1797484737547206656 |
|---|---|
| author | Stanton R. Price Derek T. Anderson Timothy C. Havens Steven R. Price |
| author_facet | Stanton R. Price Derek T. Anderson Timothy C. Havens Steven R. Price |
| author_sort | Stanton R. Price |
| collection | DOAJ |
| description | Kernel theory is a demonstrated tool that has made its way into nearly all areas of machine learning. However, a serious limitation of kernel methods is knowing which kernel is needed in practice. <i>Multiple kernel learning</i> (MKL) is an attempt to learn a new tailored kernel through the aggregation of a set of valid known kernels. There are generally three approaches to MKL: fixed rules, heuristics, and optimization. Optimization is the most popular; however, a shortcoming of most optimization approaches is that they are tightly coupled with the underlying objective function and overfitting occurs. Herein, we take a different approach to MKL. Specifically, we explore different divergence measures on the values in the kernel matrices and in the <i>reproducing kernel Hilbert space</i> (RKHS). Experiments on benchmark datasets and a computer vision feature learning task in explosive hazard detection demonstrate the effectiveness and generalizability of our proposed methods. |
| first_indexed | 2024-03-09T23:08:44Z |
| format | Article |
| id | doaj.art-50aab262b11a4f52833fed54b861412e |
| institution | Directory Open Access Journal |
| issn | 2227-7390 |
| language | English |
| last_indexed | 2024-03-09T23:08:44Z |
| publishDate | 2022-06-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Mathematics |
| spelling | doaj.art-50aab262b11a4f52833fed54b861412e2023-11-23T17:48:30ZengMDPI AGMathematics2227-73902022-06-011012202610.3390/math10122026Kernel Matrix-Based Heuristic Multiple Kernel LearningStanton R. Price0Derek T. Anderson1Timothy C. Havens2Steven R. Price3U.S. Army Engineer Research and Development Center, Geotechnical and Structures Laboratory, Vicksburg, MS 39180, USADepartment of Electrical Engineering and Computer Science, University of Missouri, Columbia, MO 65211, USADepartment of Electrical Engineering and Computer Science, College of Computing, Michigan Technological University, Houghton, MI 49931, USAU.S. Army Engineer Research and Development Center, Geotechnical and Structures Laboratory, Vicksburg, MS 39180, USAKernel theory is a demonstrated tool that has made its way into nearly all areas of machine learning. However, a serious limitation of kernel methods is knowing which kernel is needed in practice. <i>Multiple kernel learning</i> (MKL) is an attempt to learn a new tailored kernel through the aggregation of a set of valid known kernels. There are generally three approaches to MKL: fixed rules, heuristics, and optimization. Optimization is the most popular; however, a shortcoming of most optimization approaches is that they are tightly coupled with the underlying objective function and overfitting occurs. Herein, we take a different approach to MKL. Specifically, we explore different divergence measures on the values in the kernel matrices and in the <i>reproducing kernel Hilbert space</i> (RKHS). Experiments on benchmark datasets and a computer vision feature learning task in explosive hazard detection demonstrate the effectiveness and generalizability of our proposed methods.https://www.mdpi.com/2227-7390/10/12/2026<i>multiple kernel learning</i>divergence measuresheuristicsSVM |
| spellingShingle | Stanton R. Price Derek T. Anderson Timothy C. Havens Steven R. Price Kernel Matrix-Based Heuristic Multiple Kernel Learning Mathematics <i>multiple kernel learning</i> divergence measures heuristics SVM |
| title | Kernel Matrix-Based Heuristic Multiple Kernel Learning |
| title_full | Kernel Matrix-Based Heuristic Multiple Kernel Learning |
| title_fullStr | Kernel Matrix-Based Heuristic Multiple Kernel Learning |
| title_full_unstemmed | Kernel Matrix-Based Heuristic Multiple Kernel Learning |
| title_short | Kernel Matrix-Based Heuristic Multiple Kernel Learning |
| title_sort | kernel matrix based heuristic multiple kernel learning |
| topic | <i>multiple kernel learning</i> divergence measures heuristics SVM |
| url | https://www.mdpi.com/2227-7390/10/12/2026 |
| work_keys_str_mv | AT stantonrprice kernelmatrixbasedheuristicmultiplekernellearning AT derektanderson kernelmatrixbasedheuristicmultiplekernellearning AT timothychavens kernelmatrixbasedheuristicmultiplekernellearning AT stevenrprice kernelmatrixbasedheuristicmultiplekernellearning |