文献类型：期刊文献

英文题名：A new improved maximal relevance and minimal redundancy method based on feature subset

作者：Xie, Shanshan[1] Zhang, Yan[2] Lv, Danjv[1] Chen, Xu[1] Lu, Jing[1] Liu, Jiang[3]

第一作者：Xie, Shanshan

通信作者：Zhang, Y[1]

机构：[1]Southwest Forestry Univ, Coll Big Data & Intelligent Engn, Kunming 650224, Yunnan, Peoples R China;[2]Southwest Forestry Univ, Coll Math & Phys, Kunming 650224, Yunnan, Peoples R China;[3]Chinese Acad Forestry, Res Inst Forestry Policy & Informat, Beijing 100091, Peoples R China

年份：0

外文期刊名：JOURNAL OF SUPERCOMPUTING

收录：;EI(收录号：20223612686823);Scopus(收录号：2-s2.0-85137121458);WOS:【SCI-EXPANDED(收录号:WOS:000847608600001)】；

基金：This study was supported by the National Natural Science Foundation of China under Grant No. 61462078 and under Grant No. 31860332, the Yunnan Provincial Science and Technology Department under Grant No. 202002AA10007, the Yunnan Provincial Department of Education under Grant No. 2022Y558.

语种：英文

外文关键词：mRMR; Feature subset; Feature selection; ImRMR; Sequence forward search

摘要：Feature selection plays a very significant role for the success of pattern recognition and data mining. Based on the maximal relevance and minimal redundancy (mRMR) method, combined with feature subset, this paper proposes an improved maximal relevance and minimal redundancy (ImRMR) feature selection method based on feature subset. In ImRMR, the Pearson correlation coefficient and mutual information are first used to measure the relevance of a single feature to the sample category, and a factor is introduced to adjust the weights of the two measurement criteria. And an equal grouping method is exploited to generate candidate feature subsets according to the ranking features. Then, the relevance and redundancy of candidate feature subsets are calculated and the ordered sequence of these feature subsets is gained by incremental search method. Finally, the final optimal feature subset is obtained from these feature subsets by combining the sequence forward search method and the classification learning algorithm. Experiments are conducted on seven datasets. The results show that ImRMR can effectively remove irrelevant and redundant features, which can not only reduce the dimension of sample features and time of model training and prediction, but also improve the classification performance.