A user review data-driven supplier ranking model using aspect-based sentiment analysis and fuzzy theory
Bingli Sun,Xiao Song,Wenxin Li,Lu Liu,Guanghong Gong,Yan Zhao
DOI: https://doi.org/10.1016/j.engappai.2023.107224
IF: 8
2023-10-15
Engineering Applications of Artificial Intelligence
Abstract:Background: The supplier selection problem is a sophisticated decision-making process that involves evaluating multiple factors. While previous research has primarily focused on objective attributes, such as supplier qualifications, product quality, and price, the subjective opinions of users have often been overlooked. However, with the growing importance of user reviews and sentiment analysis in e-commerce, incorporating users' opinions on supplier products can provide valuable insights. Purpose: This study aims to address the limitations of existing supplier selection approaches by proposing a comprehensive framework that integrates aspect-level sentiment analysis and a fuzzy multi-attribute decision model. The goal is to enhance the decision-making process by considering both objective attributes and subjective opinions. Methods: To achieve this, we develop a novel convolutional neural network (CNN) model with a gating mechanism to perform aspect-level sentiment analysis. Furthermore, we propose a fuzzy multi-attribute decision model that combines the predefined sentiment aspects with traditional evaluation criteria. The model is applied to a dataset specifically designed for automotive component supplier selection. Results: Experimental results demonstrate the superior performance of our approach compared to existing methods and datasets. A case study demonstrates the combination of aspect-level sentiment analysis and the fuzzy decision model allows for a more comprehensive evaluation of suppliers. Conclusion: By integrating aspect-level sentiment analysis and the fuzzy multi-attribute decision model, our proposed framework offers a novel perspective on supplier selection problems. The results highlight the feasibility and superiority of our approach, providing valuable insights for management in making informed decisions. This research contributes to the fields of supplier selection, sentiment analysis, and decision-making, with potential applications in various industries beyond the automotive sector.
automation & control systems,computer science, artificial intelligence,engineering, electrical & electronic, multidisciplinary