Title: A multi-view approach to multi-criteria decision making

Authors: Francisco J. dos Santos; André L.V. Coelho

Addresses: Graduate Program in Applied Informatics, University of Fortaleza, 60811-905, Fortaleza, CE, Brazil ' Alana AI, Tabapuã St., 627, Room 13, 04533-903, Itaim Bibi, SP, Brazil

Abstract: In this paper, we investigate a new approach to multi-criteria decision making (MCDM) centred upon the application of canonical correlation analysis (CCA) to distinct groups of judgement criteria. By resorting to MV-MCDM (multi-view multi-criteria decision making), one can estimate reliable values for criteria weights via CCA for multi-view multi-criteria problems; reduce the dimensionality of the decision matrix by considering only one of the available views; and easily extend well-known MCDM methods, such as simple additive weighting (SAW) and technique for order of preferences by similarity to ideal solution (TOPSIS). MV-MCDM also allows the adoption of different aggregation methods (such as the Choquet integral and a new heuristic based on radar charts) to generate the overall scores of the alternatives. A numerical example with the multi-view versions of SAW and TOPSIS demonstrates the applicability of the novel approach.

Keywords: multi-criteria decision making; MCDM; multi-view canonical correlation analysis; CCA; TOPSIS; simple additive weighting; SAW; Choquet integral.

DOI: 10.1504/IJIDS.2023.129655

International Journal of Information and Decision Sciences, 2023 Vol.15 No.1, pp.1 - 26

Received: 26 Sep 2020
Accepted: 27 Dec 2020
Published online: 20 Mar 2023 *