EURO-Online login
- New to EURO? Create an account
- I forgot my username and/or my password.
- Help with cookies
(important for IE8 users)
3378. A Comprehensive Framework for Multi-Criteria Sorting Model Selection Based on Experimental Analysis of Model Expressiveness and Robustness of Recommendation
Invited abstract in session TA-44: Robustness analysis in MCDA 2, stream Multiple Criteria Decision Analysis.
Tuesday, 8:30-10:00Room: 20 (building: 324)
Authors (first author is the speaker)
| 1. | Michał Wójcik
|
| Institute of Computing Science, Poznan University of Technology | |
| 2. | Milosz Kadzinski
|
| Institute of Computing Science, Poznan University of Technology | |
| 3. | Moha Ghaderi
|
| Economics and Business, Pompeu Fabra University |
Abstract
We present research results on the expressiveness of models and the robustness of recommendations for various approaches to solving the value-driven threshold-based multiple criteria sorting problem. The basic model we consider is UTADIS, which assumes the evaluation of alternatives based on monotonic, preferentially independent criteria. We also present several extensions to the model. Some of them relax monotonicity constraints, and others introduce interactions between criteria.The considerations focus on the preference disaggregation paradigm, reproducing the Decision Maker's preferences, expressed as assignment examples. We present the results of the performed experimental analysis for six different approaches, focusing on the aspect of expressiveness, which is the ability to reproduce indirectly expressed preferences, and robustness, which determines the stability of the recommended assignments, tested based on all models consistent with the provided recommendations, using Robust Ordinal Regression and Stochastic Multicriteria Acceptability Analysis. The analysis results were presented and discussed based on seven quality measures, divided into the considered problem dimensions, such as the number of classes, criteria, characteristic points in marginal functions, and the richness of knowledge provided by the DM. The analysis captured the trade-off between the considered aspects, indicating that more flexible approaches with richer representation provide more expressive
Keywords
- Decision Analysis
- Decision Support Systems
- Multi-Objective Decision Making
Status: accepted
Back to the list of papers