Deceptive Financial Reporting Detection: A Hierarchical Clustering Approach Based on Linguistic Features

Abstract

Linguistic techniques have shown to be a promising way to uncover financial reports manipulations. In this paper, a linguistic features-based hierarchical clustering approach is proposed to detect deceptive financial reporting. The approach contains three steps: represents the textual data of financial reports, selects the distance function and linkage, performs hierarchical clustering and finds the deceptive reports. To verify the effectiveness of the proposed approach, five firm's annual reports are chosen as the detecting targets. The experiments show that the proposed approach gives superior results. It indicates that our findings have implications in assessing the likelihood of deceptive financial reporting.