Machine Learning Tool for Automatic ASA Detection

The application of machine learning tools has shown its advantages in medical aided decision. This paper presents the implementation of three supervised learning algorithms: the C4.5 decision tree classi er, the Support Vector Machines (SVM) and the Multilayer Perceptron MLP's in MATLAB environment, on the preoperative assessment database. The classi cation models were trained using a new database collected from 898 patients, each of whom being represented by 17 features and included in one among 4 classes. The patients in this database were selected from di erent private clinics and hospitals of western Algeria. In this paper, the proposed system is devoted to the automatic detection of some typical features corresponding to the American Society of Anesthesiolo-gists sores (ASA scores). These characteristics are widely used by all Doctors Specialized in Anesthesia (DSA's) in pre-anesthesia examinations. Moreover, the robustness of our system was evaluated using a 10-fold cross-validation method and the results of the three proposed classi ers were compared.