Leveraging Artificial Intelligence for Automated Transaction Mapping to Accounting Standards

EasyChair Preprint 15077

Abstract

For companies, adhering to industry standards, regulations, and laws is essential. This includes the requirement to file annual financial statements using accounting standards issued by relevant government agencies. However, companies often document their transactions with various non-standard descriptions, leading to discrepancies between these transactions and the accounting standards. This project aims to address these discrepancies by proposing an automated tool that uses Artificial Intelligence (AI) and Machine Learning (ML) models to align company transactions with accounting standards. Eleven AI/ML models, including Random Forest (RF) and Decision Tree Classifier (DTC), were developed and trained using historical transactions. The evaluation results indicate that while both RF and DTC models have identical training accuracies, RF performs slightly better on the test set. Consequently, RF was chosen for the automated tool.

Keyphrases: Artificial Intelligence, Random Forest, accounting standards, accounting transactions, automated mapping

@booklet{EasyChair:15077,
  author    = {Linda William and Renyu Lu and Ester Goh and Woelly William},
  title     = {Leveraging Artificial Intelligence for Automated Transaction Mapping to Accounting Standards},
  howpublished = {EasyChair Preprint 15077},
  year      = {EasyChair, 2024}}