A Multi-granular Group Decision-Making Approach Integrating Sentiment Analysis for Clustering According to Positivity

Group Decision-Making methods facilitate experts’ collective selection among limited options, which is crucial in improving decisional efficiency. With increasing access to the Internet, both the number of participants and the number of alternatives have increased, diversifying the range of evaluation among experts. Nevertheless, this complexity represents a challenge for traditional Group Decision-Making methods, such as Multi-granular Group Decision-Making, initially designed to handle differences in experts’ decisions as they employ different valuation measures. These methods are now adapted to modern complexities, where experts express their opinions through comments, challenging automated systems that rely on precise languages. Applying Natural Language Processing, especially sentiment analysis, improves the understanding of comments by classifying experts into positive and negative groups. This approach consolidates group evaluations, which are relevant for reaching a consensus and optimising the effectiveness of decisions in dynamic environments.