Text classification using genetic algorithm oriented latent semantic features
作者：
Highlights：
• Genetic algorithm oriented latent semantic features are proposed.
• The proposed approach consists of feature selection and transformation stages.
• Genetic algorithms are employed in the selection of appropriate singular vectors.
• Singular vectors are not limited to the ones with largest singular values.
• The proposed approach outperforms standard LSI and feature selection methods.
摘要
•Genetic algorithm oriented latent semantic features are proposed.•The proposed approach consists of feature selection and transformation stages.•Genetic algorithms are employed in the selection of appropriate singular vectors.•Singular vectors are not limited to the ones with largest singular values.•The proposed approach outperforms standard LSI and feature selection methods.
论文关键词：Feature selection,Genetic algorithm,Latent semantic indexing,Text classification
论文评审过程：Available online 13 April 2014.
论文官网地址：https://doi.org/10.1016/j.eswa.2014.03.041
