A new preference disaggregation TOPSIS approach applied to sort corporate bonds based on financial statements and expert's assessment

This paper presents a new version of the TOPSIS method for sorting problems. The proposed method, called Preference Disaggregation on Technique for Order of Preference by Similarity to Ideal Solution - Sort (PDTOPSIS-Sort), is based on nonlinear programming for inferring parameters and uses expert's holistic evaluations. The existing TOPSIS-Sort method demands a significant number of parameters from the expert, including the definition of boundary profiles and weights. The proposed method contributes to the literature by relieving the demand for cognitive effort observed in prior methods. Instead of providing boundary profiles for the limit between every two consecutive classes, the expert provides decision examples. In addition, the specification of weights is not required. A numerical validation of PDTOPSIS-Sort was undertaken based on results previously obtained from the literature for TOPSIS-Sort, which has been presented in detail as supplementary material. In addition, the first analysis of Brazilian corporate bonds supported by an MCDM/A model is presented. To do so, data were collected from the financial statements published by the issuers of these bonds. In total, the method proposed classified 50 debentures and the results were consistent with the preferences of the decision-maker, an investment-banking expert. (C) 2020 Elsevier Ltd. All rights reserved.