Supporting group cruise decisions with online collective wisdom: An integrated approach combining review helpfulness analysis and consensus in social networks
Feixia Ji,Jian Wu,Francisco Chiclana,Qi Sun,Changyong Liang,Enrique Herrera-Viedma
DOI: https://doi.org/10.1016/j.ipm.2024.103936
IF: 7.466
2024-10-31
Information Processing & Management
Abstract:Online cruise reviews provide valuable insights for group cruise evaluations, but the vast quantity and varied quality of reviews pose significant challenges. Further complications arise from the intricate social network structures and divergent preferences among decision-makers (DMs), impeding consensus on cruise evaluations. This paper proposes a novel two-stage methodology to address these issues. In the first stage, an inherent helpfulness level–personalized helpfulness level (IHL–PHL) model is devised to evaluate review helpfulness, considering not only inherent review quality but also personalized relevance to the specific DMs' contexts. Leveraging deep learning techniques like Sentence-BERT and neural networks, the IHL–PHL model identifies high-quality, highly relevant reviews tailored as decision support data for DMs with limited cruise familiarity. The second stage facilitates consensus among DMs within overlapping social trust networks. A binary trust propagation method is developed to optimize trust propagation across overlapping communities by strategically selecting key bridging nodes. Building upon this, a constrained maximum consensus model is proposed to maximize group agreement while limiting preference adjustments based on trust-constrained willingness, thereby preventing inefficient iterations. The proposed model is verified with a dataset of 7481 reviews for four cruise alternatives. Finally, some comparisons, theoretical and practical implications are provided. Overall, this paper offers a comprehensive methodology for real-world group cruise evaluation, using online reviews from platforms like CruiseCritic as a form of collective wisdom to support decision-making.
computer science, information systems,information science & library science