中文摘要：

目的：探讨基因海量时间序列信号的归类方法。方法：将小波多尺度分析引入，在多个尺度上进行聚类，并结合FCM得到新的聚类算法。结果：应用新的聚类方法，对小脑组织的一组基因芯片时间序列信号进行分析，通过分类结果对照发现，各类中的大多数基因生物学意义接近。结论：此改进的聚类方法是有效的，是一种新的思路。

英文摘要：

Objective in order to establish new classification method on huge data of time series of gene expression. Methods Multiscale method was applied in FCM clustering algorithm and clustering results in different scale were got. Results the new method was applied in gene expression of brain. The classification results comparison found that majority of genes in various types have close biological significance. Conclusion it is effective to gene clustering where use an improved FCM algorithm, it can give guidance in biology gene clustering and explaining gene function.