The best of two worlds: Balancing model strength and comprehensibility in business failure prediction using spline-rule ensembles

Abstract

International audienceNumerous organizations and companies rely upon business failure prediction to assess and minimize the risk of initiating business relationships with partners, clients, debtors or suppliers. Advances in research on business failure prediction have been largely dominated by algorithmic development and comparisons led by a focus on improvements in model accuracy. In this context, ensemble learning has recently emerged as a class of particularly well-performing methods, albeit often at the expense of increased model complexity. However, in practice, model choice is rarely based on predictive performance alone. Models should be comprehensible and justifiable to assess their compliance with common sense and business logic, and guarantee their acceptance throughout the organization. A promising ensemble classification algorithm that has been shown to reconcile performance and comprehensibility are rule ensembles. In this study, an extension entitled spline-rule ensembles is introduced and validated in the domain of business failure prediction. Spline-rule ensemble complement rules and linear terms found in conventional rule ensembles with smooth functions with the aim of better accommodating nonlinear simple effects of individual features on business failure. Experiments on a large selection of 21 datasets of European companies in various sectors and countries (i) demonstrate superior predictive performance of spline-rule ensembles over a set of well-established yet powerful benchmark methods, (ii) show the superiority of spline-rule ensembles over conventional rule ensembles and thus demonstrate the value of the incorporation of smoothing splines, (iii) investigate the impact of alternative term regularization procedures and (iv) illustrate the comprehensibility of the resulting models through a case study. In particular, the ability of the technique to reveal the extent and the way in which predictors impact business failure, and if and how variables interact, are exemplified

Family Business Owner as a Central Figure in Customer Relationship Management

“…• company size, when moving from direct control to integrating specialist officers [51]; • industry sector, differentiating between manufacturing and services and type of product (sole or with renewal and maintenance) [52]; • type of customer-long lasting business or one-time-purchase private consumer [53].…”

Predicting customer demand for remanufactured products: A data-mining approach

“…Two techniques which tackle this task effectively are variable importance ranking (VIR) and partial dependence plot (PDP). These methods are commonly used in many domains, such as education (Masci et al, 2018), ecology (Cutler et al, 2007), business risk management (De Bock, 2017), and supply-chain finance risk prediction (Zhu, Zhou, Xie, Wang, & Nguyen, 2019). To the best of our knowledge, however, they have not yet been applied in remanufacturing and CLSC management.…”

Refining understanding of corporate failure through a topological data analysis mapping of Altman’s Z-score model

“…Early works in this area combined MDA and logistic models (Li et al, 2012), while more recent work uses a fuller suite of ML models as well (Choi et al, 2018). De Bock (2017) exposits how spline-rule ensembles can move learning beyond the linear combinations and, in so doing, address many of the non-linearities present within financial data. Unlike standard ensemble models, these rule based ensemble learners can include different candidate algorithms at each node of the decision tree, different combination rules for each node, and offer greater options for model interpretation.…”

“…Financial practitioners are keen to know which variables are important and what probability of default is associated with each level of the explanatory factors. De Bock (2017) exposition of spline-rule ensemble learning shows how low cash ratios are associated with higher failure probabilities but increasing percentages of late payments being made by the company are unsurprisingly positively linked to failure. Some factors such as the solvency ratio and return on investment are non-linear.…”