Deep learning is used to learn machines how to reconstruct journal entries. We developed basic models trained on accounting journals from 2007 to 2018 and then tested in the fiscal year 2019. Still, lots of hyperparameters need to be checked if we want to improve accuracy.
Keywords: general ledger; journal entry; bookkeeping; accounting; deep learning; variational autoencoder; long short-term memory; anomaly detection; accounting control system.
Auto106
http://ceur-ws.org/Vol-2646/05-paper.pdf