Enhancing Classification Accuracy With K-Nearest Neighbors: Optimizing Distance Metrics and Handling Unlabeled Data
DOI:
https://doi.org/10.26713/cma.v15i4.3278Keywords:
K-nearest neighbor, Non-classified data, Adaptive K-NN, OutliersAbstract
The K-Nearest Neighbors (K-NN) is a supervised machine learning algorithm, specifically within the realm of classification and regression tasks. It is a simple yet powerful method used for making predictions based on similarities between data points. The fundamental idea behind K-NN is to classify or predict a new data point's label or classify by looking at the labels of its nearest neighbors in the training dataset. This research introduces an adaptive K-NN classification approach that leverages local data characteristics to dynamically adjust neighborhood size and distance metrics. This adaptive K-NN classification approach is thoroughly examined through comparisons with \(k=3\), \(k=5\) and \(k=7\). By adjusting neighborhood size and distance metrics, the study yields nuanced performance insights. Calculated outlier-to-class distances offer valuable adaptability indications. The experimental results showcase its potential to enhance classification accuracy and adaptability in diverse data scenarios. This method contributes to the advancement of K-NN based classification techniques and provides a promising direction for improving the efficiency of data classification tasks in real-world applications.
Downloads
References
Y.-L. Cai, D. Ji and D. F. Cai, A KNN research paper classification method based on shared nearest neighbor, in: Proceedings of NTCIR-8 Workshop Meeting, June 15–18, 2010, Tokyo, Japan, pp. 336 – 340 (2010), URL: https://research.nii.ac.jp/ntcir/workshop/OnlineProceedings8/NTCIR/07-NTCIR8-PATMN-CaiY.pdf.
C.-R. Chen and U. T. Kartini, k-Nearest neighbor neural network models for very short-term global solar irradiance forecasting based on meteorological data, Energies 10(2) (2017), 186, DOI: 10.3390/en10020186.
J. Gou, H. Ma, W. Ou, S. Zeng, Y. Rao and H. Yang, A generalized mean distancebased k-nearest neighbor classifier, Expert Systems with Applications 115 (2019), 356 – 372, DOI: 10.1016/j.eswa.2018.08.021.
J. M. Keller, M. R. Gray and J. A. Givens, A fuzzy K-nearest neighbor algorithm, IEEE Transactions on Systems, Man, and Cybernetics SMC-15(4) (2016), 580 – 585, DOI: 10.1109/TSMC.1985.6313426.
A. Lamba and D. Kumar, Survey on KNN and its variants, International Journal of Advanced Research in Computer and Communication Engineering 5(5) (2016), 430 – 435.
D. Lopez-Bernal, D. Balderas, P. Ponce and A. Molina, Education 4.0: Teaching the basics of KNN, LDA and simple perceptron algorithms for binary classification problems, Future Internet 13 (2021), 193, DOI: 10.3390/fi13080193.
Z. Pan, Y. Wang and Y. Pan, A new locally adaptive k-nearest neighbor algorithm based on discrimination class, Knowledge-Based System 204 (2020), 106185, DOI: 10.1016/j.knosys.2020.106185.
B. Wang, X. Gan, X. Liu, B. Yu, R. Jia, L. Huang and H. Jia, A novel weighted KNN algorithm based on RSS similarity and position distance for Wi-Fi fingerprint positioning, IEEE Access 8 (2020), 30591 – 30602, DOI: 10.1109/access.2020.2973212.
T. Xiao, F. Cao, T. Li, G. Song, K. Zhou, J. Zhu and H. Wang, KNN and re-ranking models for English patent mining at NTCIR-7, in: Proceedings of the 7th NTCIR Workshop Meeting, December 16–19, 2008, Tokyo, Japan, pp. 333 – 340, URL: https://research.nii.ac.jp/ntcir/workshop/OnlineProceedings7/pdf/NTCIR7/C3/PATMN/02-NTCIR7-PATMN-XiaoT.pdf.
S. Zhang, X. Li, M. Zong, X. Zhu and D. Cheng, Learning K for KNN classification, ACM Transactions on Intelligent Systems and Technology 8 (2017), Article number 43, 1 – 19, DOI: 10.1145/2990508.
Downloads
Published
How to Cite
Issue
Section
License
Authors who publish with this journal agree to the following terms:
- Authors retain copyright and grant the journal right of first publication with the work simultaneously licensed under a CCAL that allows others to share the work with an acknowledgement of the work's authorship and initial publication in this journal.
- Authors are able to enter into separate, additional contractual arrangements for the non-exclusive distribution of the journal's published version of the work (e.g., post it to an institutional repository or publish it in a book), with an acknowledgement of its initial publication in this journal.
- Authors are permitted and encouraged to post their work online (e.g., in institutional repositories or on their website) prior to and during the submission process, as it can lead to productive exchanges, as well as earlier and greater citation of published work.
