IAHR Document Library

« Back to Library Homepage « Proceedings of the 31st IAHR World Congress (Seoul, 2005)

Cluster Validity Analysis for Flood Clustering

Author(s): Zhaocheng Hou

Linked Author(s):

Keywords: Fuzzy clustering; Flood clustering; Cluster validity analysis; Optimal cluster number

Abstract: Flood clustering is the process of determining flood patterns through clustering the selected typical floods. In clustering analysis, the number of clusters is a very important parameter, and the determination of the optimal cluster number is one of the research focuses in clustering analysis. On the basis of the Fuzzy Clustering Iterative Algorithm, a cluster validity criterion is first proposed in this paper, and the flow for specifying the optimal cluster number is given, as well. Next, the IRIS dataset and an artificial dataset are used as standards for testing the validity of the proposed criterion. Taken 13 typical floods of the Biliuhe reservoir as an example, the clustering analysis is carried out, and the validity of the proposed criterion is further testified. The genetic analysis and the clustering analysis are combined to determine the flood patterns of the main weather systems of the Biliuhe reservoir, which is a premise for carrying out the classified flood operation design with different weather systems.


Year: 2005

Copyright © 2024 International Association for Hydro-Environment Engineering and Research. All rights reserved. | Terms and Conditions