An Effective Method For Genemicoarry Diagnosis Based On Optimized Svm

The DNA microarray,also known as the DNA chip,is one of the most important inventions in the biological field of the last century.He made it possible to monitor the expression of thousands of genes at the same time.With the development of modern life,the diagnosis of disease has become a very important part of medicine.This paper research the diagnosis and prediction of disease by using DNA microarray technology.The DNA microarray dataset has the characteristics of low sample,high dimension,high redundancy and high noise,and can not be classified directly by machine learning algorithm.In this paper,we first preprocessed the dataset,then used GS,CHO's,SVM-RFE and other methods to extract features,and scored genes,and got the sorting of feature genes.Using the characteristics of the selected genes,this paper chooses the optimized support vector machine and the least squares vector machine to train the model.The model has two important parameters,and the selection of parameters has a crucial influence on the quality of the classifier.In this paper,the genetic algorithm is used to optimize the parameters,and the prediction ability of the model is improved.Prediction tests were conducted on the dataset of leukaemia,glial cancer,and diffuse large B lymphadenocarcinoma.With only 4 genes were obtained in the leukemia data,and the accuracy of prediction was 100%,which was superior to other methods.At the end of this paper,we also point out the practical significance of the feature extraction of DNA microarray,and the feature extraction has a certain guiding role in the pathological study of biology and medicine.