An Improved and Adaptive Attribute Selection Technique to Optimize Dengue Fever Prediction

Abstract

Clinical information mining is rapidly gaining popularity. Restorative information are high dimensional in nature which contains unessential elements that diminish prediction capability. Hence Attribute Optimization is required to retain only the essential features while eradicating irrelevant features. Dengue is one of the major worldwide medical related disease. It has affected millions of people throughout world while a majority of them being women. With constant upgradation of information technology and its application in healthcare domain, several cases relating to diabetes along with its symptoms are properly documented. Our study is centered on developing and implementing a new Adaptive and Dynamic Attribute Optimization algorithm to determine whether patients suffer from Dengue. Our algorithm is evaluated against some vital performance metrics and compared with other sub-modules of the proposed algorithm and traditional Genetic Algorithm. The results indicate our algorithm is more efficient and accurate in determining presence of Dengue disease. This may assist the medical experts in effective diagnosis of patients suffering from Dengue.