报告人:王声瑞教授 University of Sherbrooke, Canada
报告题目:Discovering Knowledge-Sharing Communities in Question-AnsweringForums
时间:2010年10月27日(周三)上午9:30-11:00
地点:成功楼603教室;
参加对象:数计学院教师和研究生
专家简介:
1982年毕业于河北大学数学系;1986年获法国Universite de Grenoble应用数学硕士学位,1989年获法国Institute National Polytechnique de Grenoble计算机博士学位,1990年在加拿大Laval 大学电子工程系从事博士后研究。1999年起任加拿大Sherbrooke大学教授、博士生导师。研究兴趣包括模式识别、神经网络、数据挖掘、图像处理、图像数据库、地理信息系统和导航系统等。自1992年以来每年获得加拿大国家自然科学基金个人资助,多次获得集体资助,现为加拿大和美国国家自然科学基金评委会委员。自2001年以来任加拿大21世纪联合汽车研究中心网项目主持人。在《Machine Learning》、《IEEE Transactions on Knowledge and Data Engineering》、《Pattern Recognition》等国际重要期刊和ACM SIGKDD、IEEE ICDM等数据挖掘国际顶级会议上发表论文100多篇。
报告摘要:
In this talk, I will present our recent work on identification of authoritative users and discovery of knowledge-sharing communities from an Internet question-answering forum such as Yahoo! Answers. Such communities are often seen in the business world as important means for generating value and motivating contributions. First, we estimate and then model the authority scores of participants as a mixture of gamma distributions, whose parameters are estimated using the Bayesian Information Criterion (BIC) and the Expectation-Maximization (EM) algorithm. This method allows us to automatically discriminate between authoritative and non-authoritative users. Second, we model interaction between an asker and a set of authoritative users as a type of transactional data and develop a parameter-free transactional clustering algorithm to extract user communities. Finally, we experiment our approach on data from Yahoo! Answers which represent users' activities over one full year.
版权所有 © 联系地址:福建省网络安全与密码技术重点实验室(福建师范大学) |
邮编:350007 联系电话:0591-83456050转406 |
访问量:
158962020年11月04日
|