Graduation Semester and Year
2016
Language
English
Document Type
Thesis
Degree Name
Master of Science in Electrical Engineering
Department
Electrical Engineering
First Advisor
Michael T Manry
Abstract
Nearest Neighbor algorithms are non-parametric algorithms that use distance measure techniques for classification and regressions. This thesis, first improves the traditional nearest neighbor classifier by optimizing its distance measure using a second order training algorithm. It then presents a second order method to adjust center vectors. It is shown that the distance measure weight optimization and center vector optimization, individually and together reduce the final classification error. The testing error of our algorithm is shown to be less than that of LVQ V.2.1
Keywords
Weighted nearest neighbor classifier, Weighted Euclidean distance, Center vector optimization
Disciplines
Electrical and Computer Engineering | Engineering
License
This work is licensed under a Creative Commons Attribution-NonCommercial-Share Alike 4.0 International License.
Recommended Citation
Sheth, Jugal Raju, "OPTIMAL ATTRIBUTE WEIGHTING IN A NEAREST NEIGHBOR CLASSIFIER" (2016). Electrical Engineering Theses. 390.
https://mavmatrix.uta.edu/electricaleng_theses/390
Comments
Degree granted by The University of Texas at Arlington