Year: 2026 | Month: February | Volume: 16 | Issue: 2 | Pages: 85-93
DOI: https://doi.org/10.52403/ijhsr.20260212
Trend Analysis and Forecasting of Dengue Incidence in India Using Exponential Models
Dr. S. R. Itagimath
Assistant Professor in Biostatistics, Department of Community Medicine, KMCRI Hubballi, Karnataka, India
Corresponding Author: Dr. S. R. Itagimath
ABSTRACT
Background: Dengue fever remains a major public health concern in India, characterised by periodic outbreaks and fluctuating incidence rates. Reliable forecasting models are essential for planning preventive and control strategies.
Objective: To analyse the trend of dengue cases in India and to develop a suitable exponential smoothing model for forecasting future dengue incidence in India.
Methods: A retrospective time series analysis was conducted using yearly dengue case data of India from 2015 to 2025 obtained from the National Centre for Vector Borne Diseases Control (NCVBDC) under the Ministry of Health and Family Welfare, Government of India. The exponential models were used for forecasting dengue cases, which were estimated using maximum likelihood estimation and evaluated using Stationary R – squared, Root Mean Square Error (RMSE), Mean Absolute Error (MAE) and Bayesian Information Criterion (BIC).
Results: In this study, no seasonality was found; we used exponential smoothing, which is a time-series forecasting technique in which past observations are weighted exponentially, meaning that more recent values receive greater importance than older ones. This approach is particularly useful when the data show irregular fluctuations and when recent trends are more relevant for forecasting. From the exponential model’s comparison, Holt's linear trend exponential smoothing shows the highest Stationary R-squared (0.704), lowest MAE (55,249), lowest RMSE (76,004) and lowest BIC (23.08). This indicates that Holt's linear trend exponential smoothing model captures both the level and trend better than the Simple, Brown, and Damped trend models. Hence, Holt's exponential model is the most suitable model for forecasting dengue cases in India.
Conclusion: The trend of dengue cases in India shows irregular yet increasing trends, particularly after 2020. Holt's Linear Exponential Smoothing model is most appropriate for short-term forecasting due to its ability to capture both level and trend components effectively.
Key words: Dengue, Trend Analysis, Forecasting, Exponential Smoothing, India, Time Series