摘要
传统序贯三支决策模型通常不重视样本粒子的分布情况,未能较好地融入多粒度计算思想。本文提出基于模糊认知图的序贯三支多粒度决策模型,用图结点表示等价类,以图的形式进行多粒度描述和计算,并利用启发式算法实现多层次动态决策过程中粒度的自我调整和优化。实验结果表明,相较于已有的模型,该方法能够在具有较强解释性和粒度自适应的前提下,决策效果达到甚至优于基准模型。
The traditional sequential three-way decision models usually don’t pay attention to the distribution of sample granules, and fail to integrate the idea of multi-granularity computing well. Therefore, this paper proposes a sequential three-way multi-granularity decision model based on fuzzy cognitive map. Graph nodes are used to represent equivalence classes, and multi-granularity descriptions and computing are carried out in the form of graphs. Moreover, heuristic algorithms are utilized to realize self-adjustment and optimization of granularity in the multi-level dynamic decision-making process. The experimental results show that, compared with the existing models, this method can achieve or even outperform the benchmark models on the premise of strong interpretability and adaptive granularity.
作者
陈阳
杨新
李昱洁
刘盾
CHEN Yang;YANG Xin;LI Yu-jie;LIU Dun(School of Economic Information Engineering,Southwestern Univers让y of Finance and Economics,Chengdu 611130,China;Financial Intelligence&Financial Engineering Key Laboratory of Sichuan Province,Southwestern University of Finance and Economics,Chengdu 611130,China;School of Economics and Management,Southwest Jiaotong University,Chengdu 610031,China)
出处
《模糊系统与数学》
北大核心
2021年第6期76-86,共11页
Fuzzy Systems and Mathematics
基金
国家自然科学基金资助项目(61876157)
教育部人文社会科学青年基金资助项目(20YJC630191)。
关键词
模糊认知图
序贯三支决策
等价类
粒度自适应
可解释性
Fuzzy Cognitive Map
Sequential Three-way Decision
Equivalence Classes
Granularity Adaptation
Interpretability
