Genes Selection and Tumour Classifications in Cancer Research

A common problem in microarray cancer studies is how to identify and select the most informative marker genes whose expression levels are predictive of clinical or other outcomes of interest. A major constraint however, is that the expression levels of most of these genes are often collected on relatively few samples which makes the use of classical regression methods inappropriate for genes selection and the prediction of biological samples. A number of methods have been proposed in literature some of which are characterized by procedural complexities. In this book, a novel k sequential genes selection and tumour classification (k-SS) method is proposed. The new k-SS method is very simple to implement and it competes favourably with some selected state-of-the-art methods under varying platforms. Also, a classifier-like preliminary feature selection technique that employs the cross-validated area under the ROC curve for primary gene selection is presented in this book among other results. This book should be a valuable resource for researchers in statistics, medical, pharmacological and biological sciences or anyone that is interested in data mining in science and industry.