Semi-supervized feature reduction for bio-data analysis

73 p.Cancer classification is one of the most important applications in microarray data analy-sis. Due to high dimensional natures of the microarray data, feature reduction has drawn special attention. In this project, we target to explore effective and semi-supervised fea-ture reduction methods for the gene expression data classification. Two new feature re-duction methods have been explored, analyzed and tested on several microarray datasets. One is a semi-supervised feature selection method and the other one is semi-supervised feature extraction method via label propagation.Master of Science (Signal Processing