Feature Selection for High Dimensional Data: An Evolutionary Filter Approach | Science Publications

Journal of Computer Science

A. A. Yahya¹  A. Osman¹  A. R. Ramli¹  A. Balola¹ 

关键词: Genetic algorithm;    feature selection;    high dimensional data;    filter approach;    Machine Learning (ML);    evaluation function;    proposed approach;    search algorithm;    natural language processing;    mutation operator;   

DOI  :  10.3844/jcssp.2011.800.820

学科分类：计算机科学（综合）

来源: Science Publications

【 摘 要 】

Problem statement: Feature selection is a task of crucial importance for the application ofmachine learning in various domains. In addition, the recent increase of data dimensionality poses asevere challenge to many existing feature selection approaches with respect to efficiency andeffectiveness. As an example, genetic algorithm is an effective search algorithm that lends itselfdirectly to feature selection; however this direct application is hindered by the recent increase of datadimensionality. Therefore adapting genetic algorithm to cope with the high dimensionality of the databecomes increasingly appealing. Approach: In this study, we proposed an adapted version of geneticalgorithm that can be applied for feature selection in high dimensional data. The proposed approach isbased essentially on a variable length representation scheme and a set of modified and proposedgenetic operators. To assess the effectiveness of the proposed approach, we applied it for cues phraseselection and compared its performance with a number of ranking approaches which are alwaysapplied for this task. Results and Conclusion: The results provide experimental evidences on theeffectiveness of the proposed approach for feature selection in high dimensional data.

【 授权许可】

Unknown