Bing Liu

 

Bing Liu

Associate Professor Department of Computer Science University of Illinois at Chicago (UIC) 851 S. Morgan (M/C 152) Chicago, IL 60607-7053

Tel: 1 (312) 355 1318 Fax: 1 (312) 413 0024

Past and present research interests: (after 1996-97) data mining, Web and text mining, machine learning, (before 1996-97) constraint satisfaction, route planning, and AI based scheduling. 数据挖掘研究院

Research Interests

Publications

Professional Activities

Data mining and knowledge discovery Web content mining, text mining Machine learning

By topics Chronological listing DBLP: many are not mine Google Scholar

Web Content Mining: tutorial given at WWW-2005 and WISE-2005.

Research Projects

• Learning from Positive and Unlabeled Examples
• Main feature of the problem: No labeled negative examples
• Download our LPU system - LPU and Spy-EM (S-EM) system - S-EM
• Web Content Mining, Text Mining, and Search
  • Opinion Extraction, Opinion Mining, and Feature-Based Summarization
    • We currently focus on consumer reviews on the Web. The system extracts product features that have been commented on by customers and determine their (positive and negative) opinons.
  • Web Data Extraction from Flat and Nested Data Records - automatic wrapper generation
    • Download ourMDR system; Two new systems, DEPTA and NET, but not for download yet.
  • Deep Search: Information Synthesis (or Knowledge Synthesis)
    • Given a search topic/query, discover sub-topics and related concepts and synthesize pieces of information from multiple pages to present a coherent and complete picture of the topic. Like generating a "book" on the topic.
  • Web page template detection and noise removal - page segmentation and cleaning
• Data Mining and Knowledge Discovery
  • Classification Based on Associations - Integrating Classification and Association Rule Mining
    • Download our system (CBA - Asociation rules based classification)
    • KDD-98 paper and others (KDD-99, PKDD-00, etc)
  • Association rule mining with multiple minimum supports
    • Very important for skewed data. Most real life data are skewed or highly imbalanced.
  • Interestingness: subjective interestingness, rule summary, rule query language, change mining
    • At Last, interestingness techniques have made it to the real world,. They play a central role in finding actionable knowledge in a deployed data mining system for Motorola. As a researcher on this topic for many years, this is paticularly satisfying.
  • Model Management: Managing and Analyzing Large Collections of Data Mining Models.

Teaching

• Spring 2006: CS 511 - Artificial Intelligence II
• Fall 2005: CS 583 - Data mining and text mining
• Past teaching

Professional Memberships

• Member, Association for Computing Machinery (ACM).
• Member, ACM Special Interest Group on Knowledge Discovery in Data and Data Mining (SIGKDD).
• Member, American Association of Artificial Intelligence (AAAI).
• Member, IEEE Computer Society.

---

First Draft: by Bing Liu on April 10, 2002.