H. Shu, W. Jiang and R. Yu, "Study on weak bit in Vote Count and its application in k-Nearest Neighbors Algorithm," Industrial Electronics and Applications (ICIEA), 2015 IEEE 10th Conference on, Auckland, 2015, pp. 119-122. doi: 10.1109/ICIEA.2015.7334095
Abstract:
In the Vote Count for k-Nearest Neighbors (kNN) algorithm, the quantized projection values of query and training/reference vectors are compared and counted. In this process, not all quantized projection results are reliable for bit-matching and the search result may be distorted by these unreliable bit-matching. In this paper, the concept of weak bit is introduced to identify those unreliable bits after quantization and the corresponding bit-matching comparison is not executed. Simulation results show that, when weak bit is employed, the accuracy of kNN based on Vote Count can be improved significantly.
License type:
PublisherCopyrights
Funding Info:
Description:
(c) 2015 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other users, including reprinting/ republishing this material for advertising or promotional purposes, creating new collective works for resale or redistribution to servers or lists, or reuse of any copyrighted components of this work in other works.