Intelligent classification algorithms in enhancing the performance of support vector machine
Performing feature subset and tuning support vector machine (SVM) parameter processes in parallel with the aim to increase the classification accuracy is the current research direction in SVM. Common methods associated in tuning SVM parameters will discretize the continuous value of these parameters...
| Main Authors: | , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Little Lion Scientific
2019
|
| Subjects: | |
| Online Access: | https://repo.uum.edu.my/id/eprint/27867/1/JIAIT%2097%202%202019%20664%20657.pdf |
| _version_ | 1825805704545435648 |
|---|---|
| author | Alwan, Hiba Basim Ku-Mahamud, Ku Ruhana |
| author_facet | Alwan, Hiba Basim Ku-Mahamud, Ku Ruhana |
| author_sort | Alwan, Hiba Basim |
| collection | UUM |
| description | Performing feature subset and tuning support vector machine (SVM) parameter processes in parallel with the aim to increase the classification accuracy is the current research direction in SVM. Common methods associated in tuning SVM parameters will discretize the continuous value of these parameters which
will result in low classification performance. This paper presents two intelligent algorithms that hybridized between ant colony optimization (ACO) and SVM for tuning SVM parameters and selecting feature subset without having to discretize the continuous values. This can be achieved by simultaneously executing the
selection of feature subset and tuning SVM parameters simultaneously. The algorithms are called ACOMVSVM and IACOMV-SVM. The difference between the algorithms is the size of the solution archive. The size of the archive in ACOMV is fixed while in IACOMV, the size of solution archive increases as the optimization procedure progress. Eight benchmark datasets from UCI were used in the experiments to validate the performance of the proposed algorithms. Experimental results obtained from the proposed algorithms are better when compared with other approaches in terms of classification accuracy. The average classification accuracies for the proposed ACOMV–SVM and IACOMV-SVM algorithms are 97.28 and 97.91 respectively. The work in this paper also contributes to a new direction for ACO that can deal with mixed variable ACO. |
| first_indexed | 2024-07-04T06:36:37Z |
| format | Article |
| id | uum-27867 |
| institution | Universiti Utara Malaysia |
| language | English |
| last_indexed | 2024-07-04T06:36:37Z |
| publishDate | 2019 |
| publisher | Little Lion Scientific |
| record_format | eprints |
| spelling | uum-278672020-11-10T05:59:16Z https://repo.uum.edu.my/id/eprint/27867/ Intelligent classification algorithms in enhancing the performance of support vector machine Alwan, Hiba Basim Ku-Mahamud, Ku Ruhana QA75 Electronic computers. Computer science Performing feature subset and tuning support vector machine (SVM) parameter processes in parallel with the aim to increase the classification accuracy is the current research direction in SVM. Common methods associated in tuning SVM parameters will discretize the continuous value of these parameters which will result in low classification performance. This paper presents two intelligent algorithms that hybridized between ant colony optimization (ACO) and SVM for tuning SVM parameters and selecting feature subset without having to discretize the continuous values. This can be achieved by simultaneously executing the selection of feature subset and tuning SVM parameters simultaneously. The algorithms are called ACOMVSVM and IACOMV-SVM. The difference between the algorithms is the size of the solution archive. The size of the archive in ACOMV is fixed while in IACOMV, the size of solution archive increases as the optimization procedure progress. Eight benchmark datasets from UCI were used in the experiments to validate the performance of the proposed algorithms. Experimental results obtained from the proposed algorithms are better when compared with other approaches in terms of classification accuracy. The average classification accuracies for the proposed ACOMV–SVM and IACOMV-SVM algorithms are 97.28 and 97.91 respectively. The work in this paper also contributes to a new direction for ACO that can deal with mixed variable ACO. Little Lion Scientific 2019 Article PeerReviewed application/pdf en https://repo.uum.edu.my/id/eprint/27867/1/JIAIT%2097%202%202019%20664%20657.pdf Alwan, Hiba Basim and Ku-Mahamud, Ku Ruhana (2019) Intelligent classification algorithms in enhancing the performance of support vector machine. Journal of Theoretical and Applied Information Technology, 97 (2). pp. 644-657. ISSN 19928645 http://www.jatit.org/volumes/ninetyseven2.php |
| spellingShingle | QA75 Electronic computers. Computer science Alwan, Hiba Basim Ku-Mahamud, Ku Ruhana Intelligent classification algorithms in enhancing the performance of support vector machine |
| title | Intelligent classification algorithms in enhancing the performance of support vector machine |
| title_full | Intelligent classification algorithms in enhancing the performance of support vector machine |
| title_fullStr | Intelligent classification algorithms in enhancing the performance of support vector machine |
| title_full_unstemmed | Intelligent classification algorithms in enhancing the performance of support vector machine |
| title_short | Intelligent classification algorithms in enhancing the performance of support vector machine |
| title_sort | intelligent classification algorithms in enhancing the performance of support vector machine |
| topic | QA75 Electronic computers. Computer science |
| url | https://repo.uum.edu.my/id/eprint/27867/1/JIAIT%2097%202%202019%20664%20657.pdf |
| work_keys_str_mv | AT alwanhibabasim intelligentclassificationalgorithmsinenhancingtheperformanceofsupportvectormachine AT kumahamudkuruhana intelligentclassificationalgorithmsinenhancingtheperformanceofsupportvectormachine |