Performance Evaluation of some Zero Mean Classification Functions under Unequal Misclassification Cost

  • Michael Asamoah-Boaheng Kumasi Polytechnic, Kumasi, Ghana
  • Atinuke. O. Adebanji Kwame Nkrumah University of Science and Technology

Abstract

This study considered the performance of Minimum Expected Cost of Misclassification method (MECM) and Quadratic Discriminant Function approach (QDF) for equal mean discrimination under unequal misclassification cost. Discriminant functions were derived under each of the
following misclassification cost ratios; 1: 1, 1: 2, 1: 3 and 1: 4and their error rate estimates determined using the Cross Validation (CV) and Balanced Error Rate (BER) methods. Least Mean error rates were recorded under QDF method as compared to that of MECM. The error rate estimates showed the QDF out performing the MECM in the provision of maximum separation between the two groups. Also the two classification rules were found to be sensitive to misclassification cost ratios exceeding 1:2

Published
Aug 8, 2017
How to Cite
ASAMOAH-BOAHENG, Michael; ADEBANJI, Atinuke. O.. Performance Evaluation of some Zero Mean Classification Functions under Unequal Misclassification Cost. International Conference on Applied Science and Technology Conference Proceedings, [S.l.], v. 1, n. 1, p. 292-302, aug. 2017. ISSN 2467-902X. Available at: <https://conference.kstu.edu.gh/index.php/proceedings/article/view/66>. Date accessed: 22 aug. 2019.