A note on genetic algorithms for large-scale feature selection
Pattern Recognition Letters
Mean Shift: A Robust Approach Toward Feature Space Analysis
IEEE Transactions on Pattern Analysis and Machine Intelligence
Genetic Algorithms in Search, Optimization and Machine Learning
Genetic Algorithms in Search, Optimization and Machine Learning
Improved Fast Gauss Transform and Efficient Kernel Density Estimation
ICCV '03 Proceedings of the Ninth IEEE International Conference on Computer Vision - Volume 2
IEEE Transactions on Pattern Analysis and Machine Intelligence
Pattern Classification (2nd Edition)
Pattern Classification (2nd Edition)
Improvement of ICA Based Probability Density Estimation for Pattern Recognition
ICPR '04 Proceedings of the Pattern Recognition, 17th International Conference on (ICPR'04) Volume 3 - Volume 03
A geometric approach to non-parametric density estimation
Pattern Recognition
Dimensionality reduction using genetic algorithms
IEEE Transactions on Evolutionary Computation
Hi-index | 0.00 |
This paper summarizes our new research on a unified design of the feature extraction and classification processes. We have developed a multiple class classifier, based on a genetically optimized kernel density estimator. The genetic algorithm provides the tool to solve the bandwidth matrix optimization problem. The bandwidth matrix of the kernel function plays a similar role with the matrix used in a linear feature extraction, since it weights differently the data vector components. More, the bandwidth matrix controls the smoothness of decision surfaces. Tests are made with both, the standard nonparametric k-Nearest Neighbour classifier and the genetically optimized kernel based density estimator. Results on a barley seed image feature data show the utility of the proposed approach.