Integrating Tensor-Based Data Analytics and Adaptive Prediction for Informed Decision-Making Support

• Betania Campello UNICAMP
• Leonardo Tomazeli Duarte UNICAMP

Resumo

This work proposes a novel approach to support multi-criteria decision analysis (MCDA) using tensor-based data structures and an adaptive prediction method. MCDA allows for informed decision-making involving the evaluation of different alternatives based on a set of predefined criteria. Unlike previous approaches, this methodology considers the prediction of future criteria signals rather than just consider a single-period value for the criteria. The proposed method generates a tensorial representation of the data and ranks alternatives using a MCDA method. Experimental results demonstrate that this approach outperforms existing methods, particularly in decision-making situations where future long-term consequences need to be considered. This study contributes to the development of decision support systems by providing a methodological framework that leverages the potential of signal processing and tensor-based data analysis.