Using support vector machines in predicting and classifying factors affecting preterm delivery

Abstract

Various statistical methods have been proposed in terms of predicting the outcomes of facing special factors. In the classical approaches, making the probability distribution or known probability density functions is ordinarily necessary to predict the desired outcome. However, most of the times enough information about the probability distribution of studied variables is not available to the researcher in practice. In such circumstances, we need that the predictors function well without knowing the probability distribution or probability density. It means that with the minimum assumptions, we obtain predictors with high precision. Support vector machine (SVM) is a good statistical method of prediction. The aim of this study is to compare two statistical methods, SVM and logistic regression. To that end, the data on premature infants born at Tehran Milad Hospital is collected and used.