Class Prediction Methods Applied to Microarray Data for Classification

Abstract

The use of microarray data for the analysis of gene expression has been seen to
be an important tool in biological research over the last decade. The important role
of this tool is indicated by providing patients a great benefit of predicted treatment.
There is an important question about a classification problem. The question is which
genes play an important role in the prediction of class membership? There are many
classification methods applied to microarray data to solve the classification problem.
In bioinformatics, Statistical method is addressed by using microarray data. For
example breast tissue samples could be classified as either cancerous or normal.
Microarray expression profiling has provided an exciting new technology to identify
classifiers for selection treatments to patients. Sometime in special cases, prognostic
prediction is included in class prediction. In order to predict which patient will
respond to a specified treatment we can think about two classes, including
responders and no responders. The objective may be to predict whether a new
patient is likely to respond based on the Microarray expression profile of her or his
tissue sample. That it is mean accurate prediction is of obvious value in treatment
selection. To achieve the above objectives I used many methods for class prediction
using gene expression profiles from microarray experiments. This research aims to
explain what these methods are, how these methods are applied to the microarray
dataset, analyzes the results and how feature selection is used for classification.
Furthermore, comparison of these methods and cross validation will be used to
evaluate the predictive accuracy.