关键词: 微阵列, 基因表达数据, 多分辨率分析, 支持向量机

Abstract: To search a new and effective method for feature extraction and classification based on microarray expression data.The features of gene expression are extracted by the wavelet multi-resolution analysis，the features are classified by the support vector machines and BP neural network methods.There are multi-scale feature gene expression，the maximum classification rate is 98.61％，with stable results.Multi-scale theory analysis of gene expression is a new and effective bioinformatics method，which is worth further exploration and research.

Key words: microarray, gene expression data, multi-resolution analysis（MRA）, Support Vector Machines（SVM）