An Efficient Feature Selection Method for Multiple Time Series Clinical Data Classification

Patient’s condition description consists of combination and changes of clinical measures. Conventional data processing methods and classification algorithms may reduce the prediction performance of clinical data.Inorder to improve the accuracy of clinical data outcome prediction by using feature selection method with multiple measurement support vector machine(MMSVM) classification algorithm is proposed. Most popular primary liver cancer is hepatocellular carcinoma (HCC). It stands in the fifth position in the world considering the tumour ranking. HCC can be treated by using Radiofrequency ablation (RFA). Recurrence prediction of hepatocellular carcinoma (HCC) after RFA treatment is an important task. The proposed method uses Binary krill herd method as the feature selection method for classification of clinical data.This method can be used for prediction of Hepatocellular Carcinoma (HCC) recurrence. After data processing, multiple measurement support vector machine(MMSVM) is used as classification method to predict HCC recurrence.The method classify data into two classes-1) HCC recurrence and 2) no evidence of recurrence of HCC.The performance accuracy of HCC recurrence prediction was significantly improved by using the feature selection method.