Intelligent On-line Advising with Expert System Shell

1
Intelligent On-line Advising with Expert System
Shell

• Kazimierz Kowalski (CS, California State Univ.)
• Mahabub Alam (CS, California State Univ.)
• Jozef Goetz (Math/Physics/CS, Univ. of La Verne)

Presented by Sharon HSIAO Nov.20.2006
2
agenda

• Introduction/motivation
• Expert systems in student advising
• Advising process, advising system
• Designing the system with Exsys CORVID
• conclusion

3
The NEED for an automated system

• Lot of transfer students
• Students availability is limited
• Many new faculty members
• Online education

4
Constraints of Selecting classes

• Course prerequisites
• Actual class openings
• Degree check requirements
• Students availability

5
Expert Systems in Student Advising

• Early 90s.
• PC-based expert system, Industrial and System
  Engineering Dept. Ohio University
• Similar approach, but in LISP. California State
  University Dominguez Hills
• La Crosse VP-Expert expert system shell,
  University of Wisconsin
• Exsys expert system shell, CS dept. California
  State University Dominguez Hills

6
Online advising systems

• Implementation techniques
• Client-side scripting javascript
• Server-side scripting active server pages
• Web-based interactive student advisement(Tang,Fluk
  erZeng,1999)
• Applet Java
• Michigan Technology univ.(Morrison,2003),
  Duisburgh Univ. (Werner, at al. 2003), eAdvisor
  Univ. North Carolina at Chapel Hill (Scotts,
  2003)
• Expert systems shells
• Exsys Corvid (Awad, 2002)
• Complex expert systems techniques to deliver
  highly effective expert knowledge
• Giving students advice online through expert
  systems

7
Advising Process

• Degree check requirements
• Student special interests
• Structured degree programs
• Less rigid programs
• Taking the course which is the prerequisite for
  largest number of other courses guarantees the
  shortest study time

8
Advising System
Goal automate the student advising process in
General Studies Program
Names of departments involved in selected courses
within sections Has to satisfy the number of
units required for sections and categories Has to
satisfy constraints related to the number of
departments involved Students interests,
students availability
9
Designing the System with Exsys CORVID

• Conversation in decision-making process within
  webpage
• Inference engine analyze and combine individual
  rules to solve larger problems, to determine what
  possible answers

10
Exsys CORVID

• Example rule
• Which of the following social science
• courses have you completed so far?
• Then
• SS is giving the value SS3
• And FLAG6 is given the value 1
• Eg.POL_100GEO_100
• Students availability?
• Interests?

11
conclusion

• Has been tested for 1 semester, 18 qualifiers, 64
  choices(courses), 16 numeric variables, 243 rules
• Possibilities play it real!

12
Thank you!

• Please comment!

13

• PLANNER WEB, California State University,
  Monterey Bay
• http//infoserver.csumb.edu/planner
• Advising Central, North Carolina University,
• http//www.ncsu.edu/advising_central/advisers/prep
  rof.html