Title: Discovering the optimal set of ratios to use in accounting-based models

Authors: Duarte Trigueiros; Carolina Sam

Addresses: ISTAR-IUL, University Institute of Lisbon, Av. Forças Armadas, 1600 Lisbon, Portugal ' Master of European Studies Alumni Association (MESA), Institute of European Studies of Macau (IEEM), Calcada do Gaio 6, Macau, China

Abstract: Ratios are the prime tool of financial analysis. In predictive modelling tasks, however, the use of ratios raises difficulties, the most obvious being that, in a multivariate setting, there is no guarantee that the collection of ratios eventually selected as predictors will be optimal in any sense. Using, as starting-point, a formal characterisation of cross-sectional accounting numbers, the paper shows how the multilayer perceptron can be trained to create internal representations which are an optimal set of ratios for a given modelling task. Experiments suggest that, when such ratios are utilised as predictors in well-known modelling tasks, performance improves on that reported by the extant literature.

Keywords: knowledge extraction; financial analysis; financial ratios; financial technology; fintech; accounting models; bankruptcy prediction; financial misstatement detection; earnings forecasting.

DOI: 10.1504/IJSSS.2018.092550

International Journal of Society Systems Science, 2018 Vol.10 No.2, pp.110 - 131

Received: 01 Dec 2017
Accepted: 17 Dec 2017
Published online: 24 Jun 2018 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article