An approach to tree-classifier design based on a splitting algorithm
作者:
WANG RU-YE,
期刊:
International Journal of Remote Sensing
(Taylor Available online 1986)
卷期:
Volume 7,
issue 1
页码: 89-104
ISSN:0143-1161
年代: 1986
DOI:10.1080/01431168608954663
出版商: Taylor & Francis Group
数据来源: Taylor
摘要:
The tree classifier is an effective method in statistical pattern classification. In this paper, a new tree-classifier design approach based on a splitting algorithm is proposed. This method not only has the common advantages of the tree classifier, but also enables its user to make a trade-off between accuracy and economy of computing time, and to achieve satisfactory results with a relatively short computing time. Experimental classification for test data and remote-sensing data is discussed.
点击下载:
PDF (421KB)
返 回