Application And Improved Of DNA Microarray Gene Expression Data Analysis Based On Support Vector Machine

In recent years, with the development of machine learning technology and thepopularization of medical information system in hospitals and the improvement ofmeasuring instrument technology makes the rapid growth of medical data. In the data setusing various data mining techniques to understand the relationship between a variety ofdiseases, the law of development of various diseases, summarize the therapeutic effect ofvarious treatments, and the disease diagnosis, treatment and experience is valuable to themedical research and development prospects. Gene expression data usually has a largeamount of data and the characteristics of high dimension and small sample, and SVM cansolve the problem of high dimensional data analysis, the applied to the experimental dataanalysis of SVM has great research significance. This thesis introduces the latest relevantapplication of SVM and obtained some achievements, and presents a SVM in the field ofthe most widely used a few of the latest research progress at first. Then leads out theprinciple of machine learning algorithms and give the proof. Then this thesis gives twomethods of SVM realized, introduced the use of the SMO algorithm and program flow ofgene microarray data classification and regression, and obtained good results.Finally theimplementation of SVM algorithm on Spark platform and the main code and theexperimental results are presented, the experimental results show that the speed andregression accuracy of algorithm is greatly improved in the distributed platform.Thisthesis solves the difficulty of the long running time for program and improves theefficiency of general algorithm based on support vector machine in the general computersystem.