An Intelligent Model For Detection Of Post-Operative Infections

Abstract

An effective intelligent diagnosis model is aiming to provide a comprehensive ANALYSIS to form optimal partitioning REPRESENTATION of patient data, and extracts the most significant features for each partition which raise the accuracy of diagnosis process. Optimal Clustering for support feature machine (OCSFM) is proposed to improve the feature selection in medical data classification comprises clustering, feature selection, and classification concepts which is based on fuzzy C-means, max-min, and support feature machine (SFM) models. Experiments have been conducted on database of surgical patients to detect postoperative infections. The performance of the method is evaluated using classification sensitivity, specificity, overall accuracy, and Matthew's correlation coefficient. The results show that the highest classification performance is obtained for the OCSFM model, and this is very promising compared to NaiveBayes, Linear Support Vector Machine (Linear SVM), Polykernal SVM, artificial neural network (ANN), and SFM models.