小波包分解和模糊聚类下的基因表达数据分析

Wavelet packet decomposition-based on fuzzy clustering for gene expression data

针对基因表达数据中存在的噪声对聚类分析结果准确度的影响问题,提出了一种基于小波包分解的基因表达数据模糊聚类分析方案,介绍了理论根据和算法,给出了Matlab仿真结果,并与其他方法聚类的结果进行了比较。结果表明提出的方法能够减少传统聚类方法受到噪声影响的程度,能够挖掘出基因表达数据在时间上的行为特征,对与细胞周期调控有关的基因表达数据的聚类结果划分更为准确和细致。

A wavelet packet decomposition based gene expression data clustering analysis scheme is put forward in order to reduce the noise inherent in the data.The theoretical foundation and algorithm is introduced.Results of Matlab are listed.Results of direct clustering and wavlet transform based denoise clustering are also listed for comparison.Those results show that the wavelet packet decomposion method can reduce background noise and mine the time character of gene expression so that the accuracy of cluster has been enhanced.This method is more useful for cell cycle regulated genc expression data clustering and can get more exact and detailed partitioning results.