NotesFAQContact Us
Collection
Advanced
Search Tips
ERIC Number: ED410280
Record Type: Non-Journal
Publication Date: 1997-Mar
Pages: 21
Abstractor: N/A
Reference Count: N/A
ISBN: N/A
ISSN: N/A
A Method for Selecting between Fisher's Linear Classification Functions and Least Absolute Deviation in Predictive Discriminant Analysis.
Meshbane, Alice; Morris, John D.
A method for comparing the cross-validated classification accuracy of Fisher's linear classification functions (FLCFs) and the least absolute deviation is presented under varying data conditions for the two-group classification problem. With this method, separate-group as well as total-sample proportions of current classifications can be compared for the two classification procedures. Q. McNemar's (1947) test for contrasting correlated proportions is used in statistical comparisons of separate-group and total-sample proportions. The method is illustrated with 22 real data sets. FLCFs were built on assumptions of multivariate normality, equal covariance matrices, and equal prior probabilities of group membership. Least absolute deviation (LAD) models were built using a computer subroutine that incorporates linear programming code. Use of the method and computer program demonstrated in this study will allow researchers to compare the explicit cross-validated classification hit-rate accuracy of LAD and FLCFs for any specific data set and select the procedure that yields the higher total-sample or separate-group hit rate, depending on the hit rate of interest. (Contains 1 table and 23 references.) (Author/SLD)
Publication Type: Reports - Evaluative; Speeches/Meeting Papers
Education Level: N/A
Audience: N/A
Language: English
Sponsor: N/A
Authoring Institution: N/A