Abstract
One of the most critical problems in healthcare is predicting the likelihood of hospital readmission in case of chronic diseases such as diabetes to be able to allocate necessary resources such as beds, rooms, specialists, and medical staff, for an acceptable quality of service. Unfortunately relatively few research studies in the literature attempted to tackle this problem; the majority of the research studies are concerned with predicting the likelihood of the diseases themselves. Numerous machine learning techniques are suitable for prediction. Nevertheless, there is also shortage in adequate comparative studies that specify the most suitable techniques for the prediction process. Towards this goal, this paper presents a comparative study among five common techniques in the literature for predicting the likelihood of hospital readmission in case of diabetic patients. Those techniques are logistic regression (LR) analysis, multi-layer perceptron (MLP), Naive Bayesian (NB) classifier, decision tree, and support vector machine (SVM). The comparative study is based on realistic data gathered from a number of hospitals in the United States. The comparative study revealed that SVM showed best performance, while the NB classifier and LR analysis were the worst.