The arXiv is a major open-access repository for preprints in fields like Physics, Mathematics, and Computer Science. Given the high and varying volume of daily submissions, the visibility of a preprint can depend heavily on when it is posted.

The goal of this project is to develop a short-term forecasting algorithm that predicts the daily number of preprints to be posted per category in the upcoming week. By identifying days with the lowest expected submission counts, the model would help researchers to strategically select submission dates to maximize visibility and audience reach.