Tuberculosis (TB), a fatal infectious disease caused by Mycobacterium tuberculosis , spreads through aerosol droplets from active cases. The Global TB Report 2025 estimates ~10.7 million incident cases and 1.23 million deaths worldwide. This study aimed to investigate seasonal patterns in TB incidence in India and to develop univariate time series models using monthly nationwide active TB cases from January 2017 to June 2024 obtained from NIKSHAY portal. The time series forecasting of TB was approached by both traditional and machine learning‐based models, including seasonal auto‐regressive (AR) integrated moving average (MA) (SARIMA), AR neural networks (ARNNs), hybrid model (SARIMA‐ARNN) and Bayesian structural time series (BSTS) model. The performance of the models was evaluated using accuracy measures like root mean square error (RMSE), mean absolute error (MAE) and mean absolute percentage error (MAPE). Bayesian information criterion (BIC) and Akaike information criterion (AIC) were estimated. The BSTS model achieved the highest forecasting performance, yielding the lowest RMSE (27,634.42), MAE (23,300.73) and MAPE (10.55). These findings highlight the potential of integrating Bayesian frameworks to enhance the accuracy of time series forecasting for TB cases, particularly in limited time series.
R. et al. (Thu,) studied this question.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: