Incorporating Day-of-Week Effects

Background: Infectious disease forecasting models relying on deep learning may struggle when applied to daily resolution data if the synthetic training data used to pre-train the model does not account for day-of-week reporting effects.

Question / Future Work: Future research could incorporate a day-of-week routine into the synthetic data generation process (like the one discussed in Section S3) to enable the deep learning approach to effectively forecast at a daily resolution.