期刊名称：International Journal of Advances in Soft Computing and Its Applications
印刷版ISSN：2074-8523
出版年度：2017
卷号：9
期号：2
页码：45
出版社：International Center for Scientific Research and Studies
摘要：Sparse PCA and non-linear dimensional reduction methods have been developed and studied in depth for almost two decades. Their applications are huge such as speech recognition and face recognition. However, the sparse PCA and the non-linear dimensional reduction methods have not been applied to biological network inference problem. Thus, in this paper, we propose two new non-linear dimensional reduction methods which are un-normalized Laplacian Eigenmaps algorithm and symmetric normalized Laplacian Eigenmaps algorithm and the sparse PCA algorithm and apply these three new methods to the biological network inference problem using gene expression data. Experimental results show that the combination of Laplacian Eigenmaps methods and the un-supervised learning method and the combination of the sparse PCA method and the un-supervised learning method outperform the un-supervised learning method alone in terms of accuracy performance measures.
