报告题目:Fuzzy rule-based model of High- dimensional systems
报告所属学科:管理科学与工程
报告人:Hanyu E(加拿大 阿尔伯塔大学)
报告时间:2023年11月29日 9:00-12:00
报告地点:腾讯会议 512820549
报告摘要:
In this lecture, we introduce an original relation factorization method based on fuzzy relational calculus, incorporating triangular norm relations. Initially, we provide the necessary background on triangular fuzzy norms to familiarize the audience with the concept. Subsequently, we propose a fuzzy matrix factorization method based on matrix decomposition models, significantly enhancing the model's performance and interpretability.
In our fuzzy matrix factorization model, we utilize matrix decomposition techniques to model fuzzy relations. By decomposing the fuzzy matrix, we effectively capture the latent structures in the data. This model not only improves prediction accuracy but also provides interpretability of the model results. In constructing models based on fuzzy relations, we employ factorization methods to encode the conditional parts of rules. This approach renders the conditions of the rules more interpretable and logical. By replacing traditional clustering algorithms with factorization, we avoid the concentration effect in the model, thereby enhancing the model's stability and performance. Through this lecture, our aim is to introduce participants to this original relation factorization method based on fuzzy relational calculus and factorization, demonstrating its application in fuzzy rule-based models. This will aid in enhancing the model's interpretability, predictive accuracy, and stability.
报告人简介:
Hanyu E is currently serving as a Postdoctoral Fellow in the Department of Electrical & Computer Engineering at the University of Alberta in Canada. He obtained his bachelor’s degree in mechanical design, manufacturing, and automation from Xidian University in Xi'an, China, in 2015, and later earned his Ph.D. in Software Engineering and Intelligent Systems in 2022 from the University of Alberta. His supervisor is Dr. Witold Pedrycz, the chair of the Canadian Center for Computational Intelligence.
学院地址:江苏省南京市江宁区将军大道29号
邮政编码:211106
版权所有:太阳成集团(tyc3556cc·VIP认证)官网-Ultra Platform ALL RIGHTS RESERVED 苏ICP备05070685号