Medipre: Medical Diagnosis Prediction Using Rough-Regression Approximation
摘要：Rough set and regression approximations are useful in establishing decision support system for medical diagnostic applications. However, the data elimination strategy for unclassified elements or patients in the medical diagnostic applications remains as a serious issue to be explored, especially with the aim of achieving higher prediction accuracy. This paper presents step-by-step procedure in building rough-regression approximation based on data elimination strategy. A number of data sets is used to examine our proposed approximation. The result has shown that the proposed roughregression is capable to improve the prediction accuracy if compared with the existing approximations significantly. The proposed approximation can improve the performance of medical diagnosis prediction system. Therefore, it may help inexperienced doctors and patients for preliminary diagnosis.