Estimating the Spread Scenario of Acute Diarrhea and Dengue Diseases in India: A Comparative Analysis of Statistical, Mathematical, and Deep Learning Models

Abstract This work estimates the spread scenario of acute diarrhoea and dengue diseases in India. Statistical, mathematical (compartmental), and deep learning time series models are employed to forecast the spread of these diseases. Predictions are based on the reported cases and fatalities recorded between January 1, 2011, and December 31, 2021. Ten different techniques for disease spread forecast, including Regression, Bayesian Linear Regression MultiOutputRegressor + XGBoost, SIR model, Prophet, NBEATS, Gluonts, LSTM, Seq2Seq, and ARIMA statistical model, are evaluated using mean absolute percentage error (MAPE) and root mean square error (RMSE) as evaluation parameters. The ARIMA model outperforms cases of acute diarrheal disease with an RMSE value of 317.7 and a MAPE value of 2.4. The Seq2Seq model performs better for dengue cases with an RMSE value of 399.1 and a MAPE value of 6.3. This study provides valuable insights for policymakers in developing and monitoring strategies to combat these diseases.