Currently users spend a lot of time creating and updating invoicing estimations in the financial forecast tool of the organization I work for. However, forecasts are not very accurate (14% mean average error on invoicing rate end-of-year estimations).
Therefore, in this repo I am building a machine learning model to predict the amount for every expense that will be billed in the current year. This would allow for more accurate budgeting and at the same time, free up some time for users.
The data was extracted from the financial forecasting tool of the organization. For confidentiality reasons, it is not provided in this repo.