Research On Hybrid Feature Algorithm For Microarray Gene Expression Data

DNA microarray technology is composed of biology,microelectronics,computer science and finance radio chemistry as a whole,a new technique developed in the original based on nucleic acid hybridization,has received great attention in the field of medicine and biology.In recent years,In recent years,with the rapid development of large-scale high-throughput microarray technology,it has become a reality to obtain thousands of gene expression levels in an experiment.This new technology provides convenience for gene expression data collection,accessing to a large number of reaction products of mRNA gene abundance data from an experiment for microarray gene expression data,usually called by the reaction of mRNA abundance microarray data obtained by microarray,referred to as.Since the last century since 90 s,microarray technology gradually formed and profound impact on the field of biology,it appears that the detection of gene activity will become possible,microarray analysis technology application of pathological diagnosis experiment start.Since then,after more than 20 years of continuous development,the biomedical,computer and other fields will be integrated.Today,microarray technology has become a hot topic in bioinformatics,which provides a new chapter for human beings to explore biological information.The high number of dimensions,loud noise and strong redundancy of microarray has troubled the research of genetic selection.This article discuss about an algorithm for multi-dimensional microarray data,and the detection of microarray genetic expression data according to signal to Noise Ratio method,the advantage of Lasso method,Filter method,and Wrapper method.This article first review some method used in microarray data analysis,followed by the review of method in integrated systematic learning.Lastly,this article will discuss the application of new integrated techniques on microarray data.The main research are as below.(1)Introduction to microarray algorithm,and the application of characteristics of Lasso method,Filter method,and Wrapper method in microarray.Finally,the main contents of this paper are expounded.(2)The microarray data gathered through a single experiment may not be in an ideal state.It also affect the efficiency and generalization ability of classifier.The integration of multiple method used experiment is able to increase the generalization ability and the efficiency of the experiment data,and it is getting closer to the analysis ability of genetic expression data.When dealing genetic microarray data that is of high redundancy,it is impossible to conduct analysis using the traditional method.This paper presents a method combining Relief algorithm and particle swarm optimization(Relief-PSO)hybrid feature selection method,the detection algorithm for data classification results,integrated algorithm produces better generalization ability.