期刊名称：International Journal of Computer Science & Technology
印刷版ISSN：2229-4333
电子版ISSN：0976-8491
出版年度：2016
卷号：7
期号：2
页码：123-127
语种：English
出版社：Ayushmaan Technologies
摘要：Gene expression data very often contain missing values. In regards to this, effective missing value estimation methods are needful though many algorithms for gene expression data analysis require a complete matrix of gene array values. In this paper, local least square imputation and weighted k-nearest neighbors(KNN) imputation are proposed to estimate missing values in the gene expression data. The proposed local least squares(LLS) imputation method gives a target gene which has missing values through a linear combination of very similar genes. The similar genes are selected by k-nearest neighbors or k coherent genes that have bigger values of Pearson Correlation coefficients. In our experiments, the proposed KNN imputation and LLS imputation method applied in e-coli bacteria dataset producing the percentages of missing values in the data.
