基因芯片能为基因组学研究提供海量的基因表达谱数据,这些数据反映了基因在不同组织细胞的不同生长发育阶段或不同生理状态下表达水平的变化[1-2]。但是,由于基因表达谱的海量性、复杂性、噪声性和高维性特点,特别是缺失数据值的大量存在[3-5],给后续的数据分析带来了较大困难,也产生了一些问题:观察到的数据与缺失数据间的差异所产生的偏倚,严 Gene chips provide massive gene expression profiling data for genomics research, and these data reflect changes in gene expression levels in different tissues and tissues at different stages of development or under different physiological states [1-2]. However, due to the massiveness, complexity, noise and high-dimensional features of gene expression profiles, especially the large number of missing data values ​​[3-5], it brings great difficulties to subsequent data analysis and also produces Some questions: The bias between the observed data and the missing data is rigorous