Clonal Selection Classification Algorithm for High-Dimensional Data
Many important problems involve classifying highdimensional data sets, which is very difficult because learning methods suffer from the curse of dimensionality. In this paper, Clonal Selection Classification Algorithm is proposed for highdimensional data. First, an automatic non-parameter uncorrelated discriminant analysis (UDA) is adopted for dimensionality reduction (DR). Due to the favorable global search and local search, Clonal Selection Algorithm (CSA) is used to design classifier. The proposed method has been extensively compared with nearest neighbor (NN) based on Principal Component Analysis and linear discrimination analysis (PCA+LDA), nearest neighbor (NN) based on UDA (UDA+NN) and FCM based on UDA (FCM+UDA) when classifying six UCI data sets and a SAR image classification problems. The results of experiment indicate the superiority of the proposed algorithm over the three other classification algorithms in term of classification accuracy and stability.
dimensionality reduction UDA Clonal Selection Algorithm (CSA) SAR image classification
Ruochen Liu Ping Zhang Licheng Jiao
Key Laboratory of Intelligent Perception and Image Understanding of Ministry of Education of China,Institute of Intelligent Information Processing, Xidian University, Xian, 710071
国际会议
无锡
英文
89-95
2010-09-17（万方平台首次上网日期，不代表论文的发表时间）