A Novel Discrete Artificial Bee Colony Algorithm for Rough Set-based Feature Selection
| Autor: | Datong Xie, Yurong Hu, Lixin Ding, Shenwen Wang |
|---|---|
| Rok vydání: | 2012 |
| Předmět: |
General Computer Science
Computer science business.industry Ant colony optimization algorithms MathematicsofComputing_NUMERICALANALYSIS Feature selection Pattern recognition Tabu search Artificial bee colony algorithm Local optimum Feature (computer vision) Rough set Artificial intelligence business Premature convergence |
| Zdroj: | International Journal of Advancements in Computing Technology. 4:295-305 |
| ISSN: | 2233-9337 2005-8039 |
| DOI: | 10.4156/ijact.vol4.issue6.34 |
| Popis: | Feature selection plays an important role in the fields of pattern recognition, data mining and machine learning. Rough set method is one of effective methods for feature selection, which can preserve the meaning of the features. Presently ant colony optimization (ACO) has been successfully applied to rough set-based feature selection, however, it has the limitations of many control parameters, premature convergence and stagnation. In this paper, we propose a novel discrete artificial bee colony algorithm to overcome these drawbacks. In our algorithm, a search mechanism centered with feature core is introduced to avoid blindly searching, in which the values of corresponding bits of these features in feature core are unchanged during the iterative process. Meanwhile, Tabu search is adopted to avoid falling into local optima for every employed bee who can not select the solution of the Tabu list in each generation. In this work, our algorithm is applied to eight datasets from UCI database and the results produced by our algorithm and some rough set-based feature selection algorithms have been compared. The results show that our algorithm outperforms the other algorithms. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|