基于多目标的双聚类算法

A Bi-clustering Algorithm Based on Multi-objective

双聚类算法是一种同时对行和列聚类的算法,在基因表达数据分析领域多有应用。传统基因表达数据分析的双聚类算法多是单目标,同时筛选既有一定数量又有显著关系的基因的多目标挖掘又往往忽略负关联模式。但是正关联和负关联信息都有重要的生物学意义。通过在多目标中加入可以包含负关联信息的新目标函数可以在约束聚类解质量的同时包含负关联信息。

Bi-clustering algorithm is an algorithm which can simultaneously cluster row data and column data and has been widely applied in analyzing gene expression data. Traditional bi-clustering algorithms for gene expression data analysis are singleobjective, and thus cannot sort out those correlated genes that satisfy multi-objectives, such as size and significance. Further,negative associations are often disregarded by current multi-objective bi-clustering algorithms. However, both positive and negative associations are thus important in biology. Multi-objective bi-clustering algorithm can get negative information when it constraint the quality of cluster solutions by adding a new objective function which can retain negative information.