A High Performance Classifier by Dimensional Tree Based Dual-kNN
- 722 Downloads
The k-Nearest Neighbors algorithm is a highly effective method for many application areas. Conceptually the other good properties are its simplicity and easy to understand. However, according to the measurement of the performance of an algorithm based on three considerations (simplicity, processing time, and prediction power), the k-NN algorithm lacks the high-speed computation and maintenance of high accuracy for different k values. The k-Nearest Neighbors algorithm is still under the influence of varying k values. Besides, the prediction accuracy fades away whenever k approaches larger values. To overcome these issues, this paper introduces a kd-tree based dual-kNN approach that concentrates on two properties to keep up the classification accuracy at different k values and upgrade processing time performance. By conducting experiments on real data sets and comparing this algorithm with two other algorithms (dual-kNN and normal-kNN), it was experimentally confirmed that the kd-tree based dual-kNN is a more effective and robust approach for classification than pure dual-kNN and normal k-NN.
KeywordsDual-kNN kd-tree based dual-kNN k-NN Robustness
- 3.Aung, S.S., Nagayama, I., Tamaki, S.: Plurality rule–based density and correlation coefficient–based clustering for K-NN. IEIE Trans. Smart Process. Comput. 6(3), 183–192 (2017)Google Scholar
- 4.Merry, B., Gain, J., Marais, P.: Accelerating kd-tree searches for all k-nearest neighbors. In: European Association for Computer Graphics, Department of Computer Science, University of Cape Town, January 2013Google Scholar
- 6.Mitchell, T.M.: Machine Learning. McGraw-Hill Companies, New York (1997). Carnegie Mellon UniversityGoogle Scholar
- 8.Blake, C., Keogh, E., Merz, C.J.: UCI repository of machine learning databases. University of California, Department of Information and Computer Science, Irvine, CA (1998). http://www.ics.uci.edu/mlearn/MLRepository.html