题目：Leveraging User Feedback for Improving Data Quality

  报告人：陈晋川

  时间：7月5日下午14点

  地点：信息楼417室

  摘要：

  Imprecise data becomes an inevitable part in many recent applications.Although there have been many studies on querying, indexing, mining,and storing uncertain data, how to improve the quality of imprecise data is still an open question. Specifically, we must deal with two problems, i.e. correctiveness and efficiency. To solve them, we propose to cleaning imprecise data based on users’ feedbacks for the query results. Most users have partial domain knowledge and are willing to improve the quality of the data bases. Hence we propose an approach to automatically remove the wrong data inconsistent with users’ knowledge. In this talk, I will briefly introduce the management of uncertain data. Then I will explain how to model users’feedbacks, how to efficiently trace back to the corresponding source data, and how to erase them