Forecasting Machine Learning Based Feature Selection for Dengue Prediction in the Early Stage

Dengue fever is rapidly becoming a prominent health concern issue affecting the tropical and sub-tropical regions. This disease can cause serious illness in humans and is spread by bleed-feeding arthropods such fleas, parasites and mosquitoes. This disease is now becoming endemic in many nations. Delayed diagnosis accompanying with unsuitable treatment may result in the risk of death. This study's objective is to inquire into the effects, signs and clinical symptoms allied with dengue. The proposed work's objective is to examine the symptoms relating to Dengue through Statistical and Analytical measures. Machine learning Classification algorithms such as Random Forest, CART and Support Vector Machine are applied to identify the significant symptoms. The symptoms from both the measures are analyzed, ranked and finally grouped as primary symptoms. The performance of the classifier is assessed using performance evaluation measures. The experimental results show that CART is more suitable for proposed prediction model for Dengue with accuracy 97%, precision 98.3%, recall 96.8% and F-measure metrics 97.5%.