Determination of Chikungunya outbreak forecasting using Gaussian Naive Bayes algorithm with bootstrap over linear regression algorithm for improved accuracy
C. Chenna Mounika, K. Thinakaran · 2025
Developing and assessing Naive Bayes and Linear Regression, two different classification algorithms, for their efficacy in predicting Chikungunya outbreaks was the main objective of this study. Using machine learning, the research sought to develop a strong prediction model that could reliably detect possible epidemics and enable prompt preventive actions. The objective of this comparison investigation was to identify the optimal algorithm for this particular application, offering significant contributions to the field of disease outbreak prediction. Two separate groups with a total sample size of 44 were used in this work to forecast Chikungunya outbreaks using a dataset of 1,500 cases. Using preset values for power (0.8), alpha (0.05), and beta (0.2), the study used the G Power test. In order to measure the predictive power of the GNB and Linear Regression models, metrics such as accuracy, Mean Absolute Error (MAE), Mean Squared Error (MSE), and Root Mean Squared Error (RMSE) were used during the building and evaluation process.The comparison study clarifies each algorithm&s;s unique benefits and drawbacks. The Chikungunya outbreaks were predicted with 98.54% accuracy by Gaussian Naive Bayes and 49.67% accuracy by Linear Regression. In comparison to Linear Regression, Naive Bayes distinguishes itself by showcasing a notably higher accuracy of 98.54% in predicting Chikungunya outbreaks, while Linear Regression achieved a comparatively lower accuracy rate. In both Linear Regression and Gaussian Naive Bayes models, we employed bootstrapping to assess their predictive performance. Linear Regression demonstrated its accuracy in predicting Chikungunya cases, while Gaussian Naive Bayes showcased its classification prowess, achieving consistent accuracy throughout the iterations. These findings provide valuable insights into model selection for Chikungunya outbreak prediction.