Improving Disease Outbreak Predictions through Hybrid Deep Learning Models in Population Health

Disease outbreaks pose significant risks to public health, necessitating accurate and timely prediction models for effective intervention. Traditional prediction methods often struggle with high-dimensional and complex datasets, leading to limited accuracy and reliability in disease forecasting. This study addresses these limitations by proposing a hybrid Deep Convolutional Neural Network-Long Short-Term Memory (DCNN-LSTM) model tailored for population health applications. The DCNN-LSTM architecture is chosen for its ability to capture spatial features through convolutional layers while leveraging LSTM layers to model temporal dependencies within health data. The study aims to improve the accuracy of disease outbreak predictions by using data from the dataset, a reliable source of health-related behavioral and demographic data. Objectives include enhancing the prediction of outbreak trends, supporting timely decision-making, and improving the understanding of health-related patterns over time. Results demonstrate that the proposed DCNN-LSTM model outperforms traditional models, such as CNN and ResNet50, by achieving 99% accuracy. These findings underscore the model's significance as a valuable tool for public health surveillance, facilitating more informed responses to potential outbreaks. The proposed framework is implemented using python. The proposed approach holds potential for advancing data-driven public health strategies and supporting broader health management initiatives.