Differential evolution based band selection in hyperspectral data classification

Authors:
Xiaobo Liu;Chao Yu;Zhihua Cai
Affiliations:
School of Computer Science, China University of Geoscience, Wuhan, Hubei, China;School of Computer Science, China University of Geoscience, Wuhan, Hubei, China;School of Computer Science, China University of Geoscience, Wuhan, Hubei, China
Venue:
ISICA'10 Proceedings of the 5th international conference on Advances in computation and intelligence
Year:
2010

Citing 3
Cited 2

Data mining: concepts and techniques

Data mining: concepts and techniques
Differential Evolution – A Simple and Efficient Heuristic for Global Optimization over Continuous Spaces

Journal of Global Optimization
Data Mining: Practical Machine Learning Tools and Techniques, Second Edition (Morgan Kaufmann Series in Data Management Systems)

Data Mining: Practical Machine Learning Tools and Techniques, Second Edition (Morgan Kaufmann Series in Data Management Systems)

Differential Evolution for automatic rule extraction from medical databases

Applied Soft Computing
Self-adaptive differential evolution for feature selection in hyperspectral image data

Applied Soft Computing

Quantified Score

Hi-index	0.00

Visualization

Abstract

Differential evolution is a new heuristic approach for global optimization method. Hyperspectral data included huge information, how to process the hyperspectral data lack of reliable and effective technique. We presented an evolutional program to select hyperspectral band, which is very fast and certainty, in order to avoid the band felling in local minima, the selected bands are contained in the entire spectral range, and have a good performance in hyperspectral data classification. The classification methods we used are very classic, which are good at the imbalance data, so that the classification accuracy are convincing, for example NBtree, Naive Bayes and J4.8. The results in classification of various kinds of minerals in uranium deposit are better than other methods, such as information entropy.