关键词: 混合遗传算法, 特征降维, 文本分类

Abstract: Genetic algorithms as an effective search tool for optimizing the overall parallel, its many applications have been accepted early. This paper makes some improvements on the foundations of former genetic algorithms which are connected with related concepts and knowledge of rough set, puts forward new adapting degree function, and focuses on genetic algorithms and rough set algorithms on feature extraction and text classification. At the same time, it compares combined algorithms with single rough set algorithms. The experiment results indicate that combined use of genetic algorithms and the rough set in feature extraction, makes drop-dimensional characteristics of the words more conducive for the classification of text data, greatly optimizes the text classification results.

Key words: mixed genetic algorithm, reduce dimension of feature, text classification