This repository contains the code and resources for my thesis project, "Disease Forecasting in a Tropical Context: A Comparative Evaluation of Model Performance and Generalizability for Dengue Fever and Influenza in Vietnam". The study evaluates and compares the performance of seven time-series forecasting models on predicting monthly incidence rates across 55 Vietnamese provinces.

📖 Project Overview

The spread of climate-sensitive infectious diseases like Dengue Fever and Influenza poses a significant public health challenge in Vietnam. This project develops a unified forecasting pipeline to assess the generalizability of various models, from simple linear baselines to more advanced deep learning architectures (like Transformers and SCINet), across two diseases with distinct transmission dynamics.

Key Features:

1.Spatio-Temporal Analysis: Evaluates model performance across 55 provinces to uncover regional heterogeneity.

2.Comprehensive Benchmarking: Compares 7 models - Linear, RLinear, MLP, LSTM, Transformer, SCINet, and PatchTST.

3.Robust Preprocessing: Implements a custom, time-aware imputation strategy to handle missing data.