Flu Prediction Using Deep Learning Models A Case Study on Influenza-Like Illness Data
Shreyas Rajendra Hole, Shreekant Salotagi, Vinothkumar Kolluru, Gaurav Kumavat, Anup A Pachghare, Y Justindhas
Abstract
Influenza-like illnesses (ILI) pose significant health and economic challenges, necessitating accurate prediction to guide timely interventions. This study explores the application of deep learning models, particularly Bidirectional LSTM networks, to predict the percentage of ILI cases based on historical data. Using a publicly available dataset, we preprocess the data by removing outliers, creating lag features, and applying hyper- parameter tuning to optimize the model. Results indicate that incorporating lag features and regularization improves model performance. The model attained a mean absolute error (MAE) of 0.077 and an R<sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup> score of 0.9743. These measurements depict forecasting accuracies. We also engage in residual analysis of the model for its predictive performance and evaluate its viability for use in public health forecasting.