Connecting campuses to promote excellence in math and science


Henry Walker

Grinnell College
Computer Science
An Expert System to Place Incoming Students in Math and CS Classes
One of my professional goals involves the inclusion of students in a variety of my scholarly activities. This talk describes one such project, which developed an expert system to use standardized tests (SATs, ACTs, Achievement Tests, Advanced Placement scores) and high school transcripts to place incoming students in mathematics classes (Precalculus, Calculus I or II, Linear Algebra and Differential Equations). The system, based on about 90 rules, automates a process done previously by faculty, clarifies rules and informal practices, and allows letters, tailored to placement data and conclusions, to be sent to all incoming students and their advisors. The system also was extended easily to place students in computer science classes (Problem Solving and Computing, CS1, CS2). Results of this student-faculty project compared favorably with placement made by faculty and led to a refereed journal publication. Work continues to use neural networks to provide another solution to this problem. While some progress has been made in this area, the current neural-network models do not achieve the same high accuracy rates as the expert-system approach.
Apply for monies for this speaker