Introduction to statistical pattern recognition (2nd ed.)
Introduction to statistical pattern recognition (2nd ed.)
Learning internal representations by error propagation
Parallel distributed processing: explorations in the microstructure of cognition, vol. 1
Pattern Classification (2nd Edition)
Pattern Classification (2nd Edition)
A local mean-based nonparametric classifier
Pattern Recognition Letters
Nearest neighbor pattern classification
IEEE Transactions on Information Theory
Expert Systems with Applications: An International Journal
Using the idea of the sparse representation to perform coarse-to-fine face recognition
Information Sciences: an International Journal
| Hi-index | 12.05 |
The k-nearest neighbor classification rule (k-NNR) is a very simple, yet powerful nonparametric classification method. As a variant of the k-NNR, a nonparametric classification method based on the local mean vector has achieved good classification performance. In pattern classification, the sample mean and sample covariance are the most important statistics related to class discriminatory information. In this paper, a new variant of the k-NNR, a nonparametric classification method based on the local mean vector and class statistics has been proposed. Not only the local information of the k nearest neighbors of the unclassified pattern in each individual class but also the global knowledge of samples in each individual class are taken into account in this new classification method. The proposed classification method is compared with the k-NNR, and the local mean-based nonparametric classification in terms of the classification error rate on the unknown patterns. Experimental results confirm the validity of this new classification approach.