Improved SVM algorithm application on cancer diagnoses
一种改进的支持向量机及其在癌症诊断中的应用

In order to improve the generalization ability,an improved SVM: SUB-SVM was presented.First,the SVM was initially trained by all the training samples,thereby producing a number of main support vectors.Second,the SVM were re-trained only by the main support vectors,thereby producing a number of sub-support vectors.Nonlinear SVM classifier which was constructed by sub-support vectors was employed to breast cancer disease diagnoses.Higher recognition rate was obtained in the prediction.