Presented at the International Conference on Computational Science (ICCS'01)
San Francisco, CA, May, 2001.
Published in Springer Verlag's Lecture Notes in Computer Science, #2074, May 2001.

 

A Comparison of Neural Networks and Classical Discriminant Analysis in Predicting Students' Mathematics Placement Examination Scores

 

S.J.Sheel, D.A.Vrooman, R.S.Renner*, S.K.Dawsey
Department of Computer Science & Department of Mathematics Coastal Carolina University Conway, SC 29528 Department of Computer Science
California State University, Chico
Chico, CA 95929-0410 USA

 

ABSTRACT
Implementing a technique that is efficient yet accurate for college student placement into the appropriate mathematics course is important. Coastal Carolina University currently groups students into entry-level mathematics courses based upon their scores on a mathematics placement examination given to incoming freshmen. This paper examines alternative placement strategies. Using multiple regression analysis, the accumulative high school grade point average, mathematics SAT, and the final grade in Algebra II were found to be the best predictors of success on a mathematics placement examination. Entry-level mathematics placement based on neural networks and discriminant analysis is contrasted with placement results from the mathematics placement test. It is found that a neural network can outperform classical discriminant analysis in correctly predicting the recommended mathematics placement. Consequently, a trained neural network can be an effective alternative to a written mathematics placement test.

Keywords: math placement, neural networks


* Present Address: Dept. of Computer Science
California State University - Chico
Chico, CA 95929-0410
E-mail: renner@ecst.csuchico.edu