Reduced Support Vector Machine Based on Nonhierarchical Clustering Techniques for Classifying Mixed Large-Scale Datasets

摘要

In recent years, the need of big and real time data makes machine learnings and statistics more popular than ever. Healthcare is the most intense in developing intelligent system using big data. Clinical trial simulations make use to provide the virtual patient population. The covariate might be continuous or categorical. The simulation conducted in this study was to accommodate the ideal large scale data according to a certain criteria. Using these data, a modeling method is used to evaluate the 2-classes classification system. The model was constructed from a very well known pattem recognition method named support vector machine. This study applied several clustering methods for reduced support vector machine (RSM) as an alternate in setting training data for constructing classification model.