Computer Engineering and Applications ›› 2022, Vol. 58 ›› Issue (21): 142-148.DOI: 10.3778/j.issn.1002-8331.2111-0245

• Pattern Recognition and Artificial Intelligence • Previous Articles     Next Articles

Evaluation of Deep Learning Recommendation System Based on Group Consensus Algorithm

LI Chuzhen, WU Xinling, YU Yuwen   

  1. 1.Department of Information and Technology, Guangdong Technology College, Zhaoqing, Guangdong 526100, China
    2.Department of Computer Science, Guangdong Polytechnic Normal University, Guangzhou 510665, China
    3.Department of Technical, Guangzhou College of Technology and Business, Foshan, Guangdong 528131, China
    4.Principal Office, Zhaoqing University, Zhaoqing, Guangdong 526100, China
  • Online:2022-11-01 Published:2022-11-01



  1. 1.广东理工学院 信息技术学院,广东 肇庆 526100 
    2.广东技术师范大学 计算机科学学院,广州 510665 
    3.广州工商学院 工学院,广东 佛山 528131
    4.肇庆学院 校长办公室,广东 肇庆 526100

Abstract: With respect to the problems that the consensus level of expert group is not high in the process of group decision-making(GDM), a GDM algorithm based on minimizing adjustment cost is constructed. Firstly, the consistency-consensus measures are defined at three levels to measure the consistency and consensus levels of fuzzy judgment matrix. Secondly, a consensus adjustment feedback mechanism based on the minimum adjustment cost is established to identify decision makers and preference values that need to adjust consensus level. Then, an optimization model with the objective function of minimizing the total cost of consensus adjustment is established to calculate the optimal adjustment parameters and the upper bound of consensus threshold for each decision-maker. Finally, a GDM algorithm based on minimizing adjustment cost is designed and its convergence is verified. The optimization experiment of deep learning recommendation system shows that the proposed GDM algorithm is more effective in terms of cost and efficiency.

Key words: group decision-making, fuzzy judgment matrix, group consensus adjustment algorithm, minimize cost

摘要: 针对群决策过程中专家群体共识水平不高的问题,构建一种基于调整成本最小化的群决策算法。该算法在三个层面定义一致性-共识性测度,用于衡量模糊判断矩阵的一致性和共识性水平;建立基于最小调整成本的共识调整反馈机制来识别需要调整的决策者和偏好值;建立以共识调整总成本最小为目标函数的最优化模型计算每个决策者的最优调整参数和共识阈值上界;设计基于调整成本最小化的群决策算法,并验证其收敛性。通过深度学习推荐系统的优选实验表明,构建的群决策算法在成本和效率方面更有效。

关键词: 群决策, 模糊判断矩阵, 群体共识调整算法, 最小化成本