Behaviors in Bookkeeping

Bookkeeping is an established process businesses need to follow in order to keep track of their financials and tax returns. When translating a machine learning model into what a customer considers to be ‘kept books’ and what a bookkeeper considers to be ‘kept books’ variability and a dependency on the customer-bookkeeper relationship come into play. Some of this can be handled by education and process changes, but other elements can be instilled by creating a logic that is applied before, during and after a machine learning model to control for the various types of error. The presentation will go over the ways we can apply logical algorithms outside of the central model to improve the central model, and create a less error-prone training set and output.