NJIT CIS 392 Text Processing, Retrieval, and Mining

1
NJIT CIS 392 Text Processing, Retrieval, and
Mining

• Instructor Y. F. Brook Wu, Ph.D.
• Material
• Van Rijsbergen Ch1
• "What do people want from IR" by Bruce Croft

2
Contact the Instructor

• In person GITC 5502, 1145-1300, Monday and
  Wednesday.
• By phone [phone]
• By e-mail (fastest way to get my attention)
  wu_at_njit.edu Please include CIS 392 in the
  subject of your e-mail.

3
Web Board for CIS 392

• Instructions of getting a web board account
  http//webboard.njit.edu/
• Class web board http//webboard.njit.edu8080/S2
  003CIS392-002/  Please check announcements and
  new messages at least three times a week and
  contribute as much as you can.
• When posting on web board, respect others and be
  considerate. 

4
Class Web Site

• CIS 392 web site
• http//web.njit.edu/wu/teaching/sp03/CIS392/CIS39
  2-Sp03.htm

5
www.turnitin.com
6
www.turnitin.com
7
Resources Journals

• From Korfhage p. xiii
• JASIST (formally JASIS)
• ACM (please visit its digital library)
• Other suggestions
• IPM (Information Processing and Management)
• Journal of Documentations

8
Resources Conferences Proceedings

• TREC Text REtrieval Conference
• http//trec.nist.gov/
• ACM/SIGIR
• http//www.sigir2001.org/
• ASIST
• http//www.asis.org
• JCDL
• http//www.acm.org/jcdl/
• CIKM
• http//www.cikm.org/

9
Exit Requirements

• You should be familiar with the followings
• Search strategies
• Text analysis
• Retrieval models
• Retrieval effectiveness measures
• Retrieval improvement techniques
• Document Warehousing Techniques
• Text Mining Applications

10
Importance of Finding the Right Information

• Personal weather, traffic, investment, etc.
• Academic prior studies for literature review.
• Medical symptoms, treatments, etc.
• Legal precedents to support arguments, etc.

11
What Makes Text Retrieval Difficult?

• Ill-understood info needs
• Ill-formed queries
• Linguistic and semantic complexities
• Word-sense ambiguity, semantic, syntactic, etc.
• and many more reasons!

12
Data Retrieval v.s. Info Retrieval (from
Rijsbergen Ch1)
13
What Do People Want from IRby W. Bruce Croft

• Source D-Lib Magazine, November 1995, or
  http//sunsite.anu.edu.au/mirrors/dlib/dlib/novemb
  er95/11croft.html
• 10. Relevance Feedback (try Google)
• 9. Information Extraction (http//www.clearforest.
  com/ )
• 8. Multimedia Retrieval (try Google image search)

14
What Do People Want from IRby W. Bruce Croft

• 7. Effective Retrieval
• 6. Routing and Filtering (try msnbc.com, or ACM
  digital library bookshelf.)
• 5. Interfaces and Browsing
• 4. "Magic" (Automatic Query Expansion)

15
What Do People Want from IRby W. Bruce Croft

• 3. Efficient, Flexible Indexing and Retrieval
• 2. Distributed IR
• 1. Integrated Solutions