Admissions Open
Admissions Open
Abstract:
In this paper, a new approach based on semi-supervised classification for the epidemic risk management is suggested. Currently, this issue represents an important medical challenge. To manage this risk, we propose a decision support system using an expertise database shared between several hospitals. The system must be able to assist medical professionals in making quick and effective decisions in such situations. A hybrid classification approach using the SVM technique with an unsupervised classification technique (K-Means) is developed to automate data analysis. Prior to the classification process, we propose a data selection method based on Formal Concept Analysis (FCA) in order to select the best combinations for MSVM binary tree classification and to reduce the patient vectors dimension.