Advisor(s) - Committee Chair
Earl Pearson, Curtis Wilkins, John Reasoner
Department of Chemistry
Master of Science
A comparison of the reliability of three pattern recognition classifiers has been made using data having a great amount of variation. The basic concepts of the Linear Learning Machine, the K Nearest Neighbor Classifier, and the Potential Function Classifier are presented. Prediction of whether a student would pass or fail freshmen Chemistry 120 was made, based on various test results. The Linear Learning Machine was found to be an unworkable classifier for this kind of data. Both the Potential Function Classifier and the K Nearest Neighbor classifier were acceptable with the Potential Function Classifier being generally a better classifier.
Chemistry | Physical Sciences and Mathematics
Wood, Larry, "A Comparison of Various Pattern Recognition Techniques" (1977). Masters Theses & Specialist Projects. Paper 3007.