An Effective Classification Model for Cancer Diagnosis Using Micro Array Gene Expression Data

Data mining algorithms are commonly used for cancer classification. Prediction models were widely used to classify cancer cells in human body. This paper focuses on finding small number of genes that can best predict the type of cancer. From the samples taken from several groups of individuals with known classes, the group to which a new individual belongs to is determined accurately. The paper uses a classical statistical technique for gene ranking and SVM classifier for gene selection and classification. The methodology was applied on two publicly available cancer databases. SVM one-against- all and one-against-one method were used with two different kernel functions and their performances are compared and promising results were achieved.