Predictive Modeling of Epidemic Diseases based on Vector-Borne Diseases using Support Vector Machine with Extreme Gradient Boost

In recent years, the globalization and progressive climatic changes lead to increase in the occurrence of Vector Borne Diseases (VBD) from endemic regions to non-endemic regions. The common forms of VBD includes malaria, chikungunya, dengue fever, yellow fever and plague. The existing methods for predicting epidemic diseases did not handle non-linear relationships and complex interactions. Therefore, this paper proposes Support Vector Machine with Extreme Gradient Boost (SVM-XGBOOST) for predicting epidemic diseases based on vector borne diseases. Initially, vector borne diseases dataset is taken which contains epidemiological data and information related to climate. Then, this data is preprocessed by using cross validation to generalize the model well and it evaluates the quality of data. After that, Least Absolute Shrinkage and Selection Operator (LASSO) or (L1Regularization) is used for selecting features which selects the relevant features automatically and handles noisy data. finally, classification is done by using SVM-XGBOOST which handles non-linear relationships and handles complex interactions. The proposed SVM-XGBOOST achieved better results in terms of accuracy (98.45%), Precision (95.50%), Recall (97.98%), and F1 score (${9 7. 2 5 \%)}$.