面向多视角决策支持的基于在线评论的群体偏好分析方法研究

This project will conduct the research on group preference analysis methods based on online reviews for multi-perspective decision support. It is refined from the latest research trends on decision theories and methods and the practice in management decision at home and abroad. It is also a new, front and exploratory research topic since it not only has extensive practical background but also has important significance on theory study and practical application. The main research contents of this project include: methods for mining the decision related topics from the online reviews; methods for multi-class sentiment analysis of the online reviews concerning decision related topics; methods for group (subgroup) consensus measurement based on the results of sentiment analysis of online reviews; methods for group (subgroup) preference aggregation and topic ranking based on multi-granularity fuzzy distribution information. Further, this project will conduct the research on constructing the visual decision analysis support system based on online reviews, and research on application of the proposed theories and methods based on the backgrounds of several typical practical decision-making problems. The aims of this project are to refine and form the problem of group preference analysis based on online reviews for multi-perspective decision support, which are with great theoretical and practical significance; propose the decision analysis methods and techniques of group preference analysis based on online reviews for multi-perspective decision support; conduct the demonstrative application research with respect to the decision support problem with practical background; and to achieve creative high-level research fruits.

面向多视角决策支持的基于在线评论的群体偏好分析方法研究，是根据目前国内外决策理论与方法研究发展动态和管理决策实践提炼出来的、具有前沿性和探索性的新课题，它不仅具有大量的实际背景，而且对此研究具有理论和实际意义。本项目的主要研究内容是：在线评论的决策相关主题挖掘方法，针对主题的在线评论多粒度情感分析方法，基于在线评论情感分析的群体（子群）偏好共识度量方法，基于多粒度模糊分布信息的群体（子群）偏好信息融合及主题排序方法；进一步地，进行基于在线评论的可视化决策支持系统原型构建的研究，依据现实若干典型的实际背景，进行理论与方法研究成果的应用研究。本项目的研究目的在于：提炼并形成科学的、有理论与实际价值的面向多视角决策支持的基于在线评论的群体偏好分析问题体系，提出并形成解决该问题的决策分析方法和技术体系，开展针对具有实际背景的决策支持的应用示范研究，取得具有创新性的高水平研究成果。

本项目研究面向多视角决策支持的基于在线评论的群体偏好分析方法，具有重要的理论意义和实际意义。通过近四年的研究工作，取得了丰硕的研究成果，其主要体现在五个方面：给出了在线评论的决策相关主题挖掘方法，提出了针对主题的在线评论多粒度情感分析方法，提出了基于在线评论情感分析的群体（子群）偏好共识度量方法，提出了基于多粒度模糊分布信息的群体（子群）偏好信息融合及主题排序方法，依据现实若干典型问题开展了应用研究。通过本项目的研究，较好地完成了预期制定的研究计划，达到了预期研究目标。与已有的同类研究成果相比，取得的一些研究成果达到了国际先进水平。针对本项目的研究，在《Tourism Management》、《Annals of Tourism Research》、《International Journal of Production Research》、《Computers & Industrial Engineering》、《IEEE Transactions on Fuzzy Systems》、《Information Sciences》、《Information Fusion》、《International Journal of Machine Learning and Cybernetics》、《系统工程理论与实践》、《系统工程学报》、《控制与决策》等国内外重要期刊上发表论文22篇，入选ESI高被引论文2篇；培养博士研究生2人和硕士研究生6人，其中1人获得辽宁省优秀硕士学位论文、1人获得东北大学优秀博士学位论文奖；获得教育部人文社会科学优秀成果奖1项。本项目研究工作的完成，对进一步发展和完善基于在线评论的群体偏好分析和决策支持的理论与方法做出了贡献。