The effectiveness of using diversity to select multiple classifier systems with varying classification thresholds
In classification applications, the goal of fusion techniques is to exploit complementary approaches and merge the information provided by these methods to provide a solution superior than any single method. Associated with choosing a methodology to fuse pattern recognition algorithms is the choice...
| Main Authors: | , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
SAGE Publishing
2018-09-01
|
| Series: | Journal of Algorithms & Computational Technology |
| Online Access: | https://doi.org/10.1177/1748301818761132 |
| _version_ | 1818460039477723136 |
|---|---|
| author | Harris K Butler Mark A Friend Kenneth W Bauer Trevor J Bihl |
| author_facet | Harris K Butler Mark A Friend Kenneth W Bauer Trevor J Bihl |
| author_sort | Harris K Butler |
| collection | DOAJ |
| description | In classification applications, the goal of fusion techniques is to exploit complementary approaches and merge the information provided by these methods to provide a solution superior than any single method. Associated with choosing a methodology to fuse pattern recognition algorithms is the choice of algorithm or algorithms to fuse. Historically, classifier ensemble accuracy has been used to select which pattern recognition algorithms are included in a multiple classifier system. More recently, research has focused on creating and evaluating diversity metrics to more effectively select ensemble members. Using a wide range of classification data sets, methodologies, and fusion techniques, current diversity research is extended by expanding classifier domains before employing fusion methodologies. The expansion is made possible with a unique classification score algorithm developed for this purpose. Correlation and linear regression techniques reveal that the relationship between diversity metrics and accuracy is tenuous and optimal ensemble selection should be based on ensemble accuracy. The strengths and weaknesses of popular diversity metrics are examined in the context of the information they provide with respect to changing classification thresholds and accuracies. |
| first_indexed | 2024-12-14T23:23:55Z |
| format | Article |
| id | doaj.art-5e21f18cc3b6490d907ca8191dd52293 |
| institution | Directory Open Access Journal |
| issn | 1748-3018 1748-3026 |
| language | English |
| last_indexed | 2024-12-14T23:23:55Z |
| publishDate | 2018-09-01 |
| publisher | SAGE Publishing |
| record_format | Article |
| series | Journal of Algorithms & Computational Technology |
| spelling | doaj.art-5e21f18cc3b6490d907ca8191dd522932022-12-21T22:43:51ZengSAGE PublishingJournal of Algorithms & Computational Technology1748-30181748-30262018-09-011210.1177/1748301818761132The effectiveness of using diversity to select multiple classifier systems with varying classification thresholdsHarris K ButlerMark A FriendKenneth W BauerTrevor J BihlIn classification applications, the goal of fusion techniques is to exploit complementary approaches and merge the information provided by these methods to provide a solution superior than any single method. Associated with choosing a methodology to fuse pattern recognition algorithms is the choice of algorithm or algorithms to fuse. Historically, classifier ensemble accuracy has been used to select which pattern recognition algorithms are included in a multiple classifier system. More recently, research has focused on creating and evaluating diversity metrics to more effectively select ensemble members. Using a wide range of classification data sets, methodologies, and fusion techniques, current diversity research is extended by expanding classifier domains before employing fusion methodologies. The expansion is made possible with a unique classification score algorithm developed for this purpose. Correlation and linear regression techniques reveal that the relationship between diversity metrics and accuracy is tenuous and optimal ensemble selection should be based on ensemble accuracy. The strengths and weaknesses of popular diversity metrics are examined in the context of the information they provide with respect to changing classification thresholds and accuracies.https://doi.org/10.1177/1748301818761132 |
| spellingShingle | Harris K Butler Mark A Friend Kenneth W Bauer Trevor J Bihl The effectiveness of using diversity to select multiple classifier systems with varying classification thresholds Journal of Algorithms & Computational Technology |
| title | The effectiveness of using diversity to select multiple classifier systems with varying classification thresholds |
| title_full | The effectiveness of using diversity to select multiple classifier systems with varying classification thresholds |
| title_fullStr | The effectiveness of using diversity to select multiple classifier systems with varying classification thresholds |
| title_full_unstemmed | The effectiveness of using diversity to select multiple classifier systems with varying classification thresholds |
| title_short | The effectiveness of using diversity to select multiple classifier systems with varying classification thresholds |
| title_sort | effectiveness of using diversity to select multiple classifier systems with varying classification thresholds |
| url | https://doi.org/10.1177/1748301818761132 |
| work_keys_str_mv | AT harriskbutler theeffectivenessofusingdiversitytoselectmultipleclassifiersystemswithvaryingclassificationthresholds AT markafriend theeffectivenessofusingdiversitytoselectmultipleclassifiersystemswithvaryingclassificationthresholds AT kennethwbauer theeffectivenessofusingdiversitytoselectmultipleclassifiersystemswithvaryingclassificationthresholds AT trevorjbihl theeffectivenessofusingdiversitytoselectmultipleclassifiersystemswithvaryingclassificationthresholds AT harriskbutler effectivenessofusingdiversitytoselectmultipleclassifiersystemswithvaryingclassificationthresholds AT markafriend effectivenessofusingdiversitytoselectmultipleclassifiersystemswithvaryingclassificationthresholds AT kennethwbauer effectivenessofusingdiversitytoselectmultipleclassifiersystemswithvaryingclassificationthresholds AT trevorjbihl effectivenessofusingdiversitytoselectmultipleclassifiersystemswithvaryingclassificationthresholds |