The prediction of profitability using accounting narratives: a variable‐precision rough set approach
Malcolm J. Beynon, Mark A. Clatworthy, Michael John Jones · Intelligent systems in accounting, finance and management/Intelligent systems in accounting, finance & management · 2004
Abstract This article utilizes a new method of data mining for the classification of companies as profitable or non‐profitable, based on a textual analysis of the respective chairman's statement. The method used is a development of the rough set theory technique, namely the variable‐precision rough sets (VPRS) model. A dichotomous sample of companies is used to construct a set of decision rules from a VPRS analysis using the textual characteristics of the chairman's statement in UK corporate annual reports. A number of descriptive measures are analyzed, including the predictive accuracy of the decision rules. Copyright © 2005 John Wiley & Sons, Ltd.