Discriminant Analysis in the Classification of Bat Species
Katie L. Gustafson
Dr. Jason E. Miller and Dr. Dean DeCock, Faculty Mentors
Discriminant Analysis is a statistical method used to classify observations to a group and predict group membership. This method has been applied in the classification of bat species with the goal of identifying endangered species. Data was collected on several species of bats and parameters detailing call frequencies, duration, and slope were extracted from electronically recorded bat calls. Preliminary results have shown overall classification rates between 70% and 85% with individual species rates ranging from 55% to 95%. We will attempt to increase these rates by adding more parameters, some of which are derived from a wavelet-based analysis of the data.
Keywords: Discriminant, bats, classification, function analysis, call sequence
Presentation Type: Oral Paper
Location: VH 1232