On Classifying Digital Accounting Documents

Advances in computing and multimedia technologies allow many accounting documents to be digitized within little cost for effective storage and access. Moreover, the amount of accounting documents is increasing rapidly, this leads to the need of developing some mechanisms to effectively manage those (semi-structured) digital accounting documents for future accounting information systems (AIS). In general, accounting documents contains such as invoices, purchase orders, checks, photographs, charts, diagrams, etc. As a result, the major functionality of future AIS is to automatically classify digital accounting documents into different categories in an effective manner. The aim of this paper is to examine flat non-hierarchical and hierarchical classification schemes for automatic classification of different types of digital accounting documents. The experimental results show that non-hierarchical classification of digital accounting documents performs better than hierarchically classifying digital accounting documents.