In this paper, a new feature selection method for neural networks is proposed using the Parzen density estimator. A new feature set is selected employing the recently published decision boundary feature selection algorithm. The selected feature set is then used to train a neural network. Using a reduced feature set, we attempt to reduce the training time of the neural network and obtain a simpler neural network, further reducing the classification time for test data. Experiments show promising results.
|Title of host publication
|IGARSS 1992 - International Geoscience and Remote Sensing Symposium
|Subtitle of host publication
|International Space Year: Space Remote Sensing
|Ruby Williamson, Tammy Stein
|Institute of Electrical and Electronics Engineers Inc.
|Number of pages
|Published - 1992
|12th Annual International Geoscience and Remote Sensing Symposium, IGARSS 1992 - Houston, United States
Duration: 1992 May 26 → 1992 May 29
|International Geoscience and Remote Sensing Symposium (IGARSS)
|12th Annual International Geoscience and Remote Sensing Symposium, IGARSS 1992
|92/5/26 → 92/5/29
Bibliographical notePublisher Copyright:
© IEEE 1992.
All Science Journal Classification (ASJC) codes
- Computer Science Applications
- General Earth and Planetary Sciences