Prediction of Epidemic Outbreaks in Specific Region using Machine Learning

Abstract

Dengue is one amongst the common infectious disease which is caused by dengue virus and transmitted to humans by mosquitoes with this many are infected in varied regions around the world per year. The reason for this virus is atmospheric conditions, which plays a vital role in the outbreak of dengue. Therefore early prediction of dengue is the key to regulate outbreak and reduces the transmission within the community. To overcome this we are using various machine learning (ML) algorithms such as Support Vector Machine (SVM), Artificial Neural Network (ANN), Random Forest tree (RF) and Decision Tree (DT) are used to predict the dengue outbreak. Prediction is done based on weather parameters like monthly wise maximum temperature, minimum temperature, average temperature, mean temperature, humidity and Precipitation which is considered as weather dataset and this weather dataset is pre- processed using label encoding function before applying into the training models. The performances of all the models are calculated based on weather dataset. After considering performance of all the models we choose random forest as a best predictor for dengue outbreak.