A new local search based hybrid genetic algorithm for feature selection

摘要: This paper presents a new hybrid genetic algorithm (HGA) for feature selection (FS), called as HGAFS. The vital aspect of this is the salient subset within reduced size. HGAFS incorporates local search operation that devised and embedded in HGA to fine-tune FS process. technique works on basis distinct informative nature input features computed by their correlation information. aim guide process so newly generated offsprings can be adjusted less correlated (distinct) consisting general special characteristics given dataset. Thus, proposed receives redundancy information among selected features. On other hand, emphasizes selecting with number using size determination scheme. We have tested our 11 real-world classification datasets having dimensions varying from 8 7129. performances been compared results existing ten well-known algorithms. It found that, produces consistently better subsets resulting accuracies.