ABSTRACT:

Dengue fever affects 400 million people annually across more than 100 tropical and subtropical countries, including regions in Africa, the Americas, and Asia. The absence of a specific drug or ready-to-use vaccine places a heavy burden on global healthcare systems. To manage outbreaks, early warning systems are essential for guiding timely intervention and control efforts. However, existing forecasting models often lack accuracy, stability, and clear interpretation. This study introduces a new model called Rennet, an ensemble wavelet neural network with exogenous factors. The model integrates climate variables verified through statistical tests to enhance prediction performance. By combining wavelet transformation with a neural network, XEWNet improves the reliability of long-term forecasts. It provides accurate dengue outbreak predictions for the regions of San Juan, Iquitos, and Ahmedabad.

OBJECTIVE:
The objective of this study is to develop a reliable and interpretable forecasting model for dengue outbreak prediction by proposing an Ensemble Wavelet Neural Network with exogenous factors (XEWNet). The model aims to enhance prediction accuracy by incorporating climate variables identified through statistical causality tests. It seeks to address the limitations of existing forecasting models by capturing complex non-linear relationships between dengue incidence and climatic factors, while maintaining computational efficiency and scalability