Combined committee machine for classifying dengue fever
Shiladitya Saha, Sankhadip Saha · 2016
The advancement of information technology has made public health management system more efficient on keeping an eye on major outbreaks like dengue fever than before. Vigorous prediction of dengue fever not only helps the management to act faster but also helps medical professionals to treat and suggest a good solution to the patients. In this context, this paper aims to develop an intelligent computing models using multilayer perceptron and support vector machine for classifying positive and negative cases of dengue. Decision taking is vital in medical field, final opinion is sometimes taken by a committee of medical experts. Keeping this fact in mind, Dempster-Shafer based classifier fusion strategy is applied here to form a committee of prediction machine for better outcome. Results show that DS theory gives as best as 96.02% with 4 MLPs and 3 SVMs while that of selected features using Fishers Score gives 96.56% accuracy.