422. A new sorting MCDM methodology with probabilistic prediction and temporal analysis
Invited abstract in session WB-8: Decision aiding under uncertainty 2, stream Multiple Criteria Decision Aiding.
Wednesday, 10:30-12:00Room: Clarendon SR 2.08
Authors (first author is the speaker)
| 1. | Caroline Mota
|
| Universidade Federal de Pernambuco, CDSID - Center for Decision Systems and Information Development | |
| 2. | Betania Campello
|
| UNICAMP |
Abstract
This study introduces a novel multitemporal and multicriteria decision-making methodology designed to sort alternatives while accounting for uncertainties in their future performance. This approach is particularly relevant for problems where the performance of alternatives is best represented by time series data, with past variations playing a significant role in the decision-making process. The methodology evaluates the temporal performance of alternatives against a set of criteria by employing a probabilistic prediction method that models data behavior and forecasts expected future values within specified confidence intervals. In contexts of high uncertainty, traditional classification methods may lose discriminative power; hence, the proposed model integrates varying levels of uncertainty to generate a valued outranking preference function for categorizing alternatives. Grounded in the principles of the PROMETHEE and FlowSort methods, the sorting procedure accommodates multiple classification scenarios. The methodology was applied to categorize five countries based on their economic and financial performance, utilizing 40 years of data from the International Monetary Fund, and the results demonstrate its efficacy and the added informational value it provides.
Keywords
- Forecasting
- Financial Modelling
- Decision Support Systems
Status: accepted
Back to the list of papers