Classification of cancerous and non-cancerous tissues of serial analysis of gene expression data through various classifiers

Cancer can be cured if detected early and it can be detected by the expression level analyzed in the suspected tissues. Serial Analysis of Gene Expression (SAGE) is a gene expression technique used to analyze the genes on the basis of expression level of the genes. The Libraries of SAGE data contained very large number of genes, considering all genes for classifying is very tedious task and is not wise thing to do. The preprocessing of the SAGE data is performed to remove the irrelevant genes by comparing the expression level of genes in normal and cancerous libraries, and the further analysis of the dataset is done considering the reduced genes. This paper compares classification techniques for classifying the cancerous and non-cancerous tissues of human brain. The Naive Bayes (NB), Linear Discriminant Analyzer (LDA), Decision Table (DT), Support Vectors Machine (SVM) and K Nearest Neighbor (KNN) classifiers have been implemented for the analysis of SAGE data. WEKA (The Waikato Environment for Knowledge Analysis) open source software which consists of a collection of machine learning algorithms for data mining used for analysis. The results obtained reveal that the K-Nearest Neighbor (KNN) and Linear Discriminant Analyzer (LDA) have given better performance over other classifies in most of the performance measures except few. Different errors measures have also been studied in this paper for SAGE data of human brain tissues. The KNN and LDA both have given significant improvement over other classifiers.