论文部分内容阅读
复杂问题的解决,单靠数据、信息和模型是不够的,还需集成专家的意见,然而专家的意见未必一致,需要对专家意见的共识情况进行综合。本文提出基于二分图网络投影算法的专家意见综合方法,并对传统的二分图边权重单一的不足进行了改进,使其减少投影过程中的信息损失,以满足复杂条件的分析需要。实证表明,改进之后的算法能够准确地反映不同专家间的意见相似度以及单个专家在群体中的独立性。最后,利用非线性映射技术,以可视化方式直观显示专家意见的共识情况。
It is not enough to rely on data, information and models alone to solve complex problems. Experts ’opinions need to be integrated. However, experts’ opinions may not be consistent and the consensus of expert opinions needs to be integrated. In this paper, an integrated approach of expert opinion based on bipartite graph network projection algorithm is proposed, and the single weakness of the traditional dichotomous graph edge weight is improved to reduce the information loss in the projection process to meet the needs of complex conditions analysis. Empirical results show that the improved algorithm can accurately reflect the similarity of opinions among different experts and the independence of individual experts in the community. Finally, the use of non-linear mapping technology to visually display the consensus of expert opinion.