报告人： Dr. Zhiwei Lin Ulster University, UK
报告题目： Concordance measure and consensus extraction in ordering sequences
时间：2016-07-25 (星期一) 11:00 ~ 12:00
地点：仓山校区成功楼603
主办：数学与计算机科学学院
参加对象：计算机专业师生及我院其他感兴趣师生。
报告摘要：Ordering sequences have been widely used to describe preferences over a set of candidates in many areas. For example, in a group decision making system, experts use ordering sequences to model their preferences of one candidate over the others. Quantification of mutual hidden pattern information in ordering sequences is key to understanding what consensus the sequences have and where the deviation arises. Concordance is one of the measures for quantifying mutual information in a set of ordering sequences. The concordance in ordering sequences refers to the degree to which ordering sequences agree on their preferences. The higher concordance the sequences have, the better decision or consensus can be reached.
This talk presents a new framework for concordance measured in the subsequence spaces, which enables us to observe how orderings are similar to each other and how each ordering sequence deviates from the others, which is significantly novel to the way of distance or similarity based measure. An efficient algorithm is proposed to calculate the concordance of N ordering sequences. The algorithm is also versatile enough to extract meaningful common patterns from the N ordering sequences and to handle ties-ordering sequences, in which symbols may have identical scores in one sequence.
专家简介：Dr. Zhiwei Lin is a lecturer in the School of Computing and Mathematics, Ulster University, UK since 2014. He received his doctoral degree from Ulster University in 2010 and both MSc and BSc from Fujian Normal University in 2005 and 2001. Zhiwei Lin has working experience in industry with SAP, British Sky broadcasting and Oracle from 2011 to 2014 before he returned to Ulster University as a lecturer. His research interests include sequence analysis, machine learning and natural language processing for text analysis.
