Assessment of some combinations of hard and fuzzy clustering techniques for regionalisation of catchments in Sefidroud basin

Cluster analysis methods are a type of well-known technique for regionalisation of catchments to perform regional flood frequency analysis. In this study, a fuzzy extension of hybrid clustering algorithms is evaluated. Self-organizing feature maps and four hierarchical clustering algorithms were use...

Full description

Saved in:
Bibliographic Details
Published inJournal of hydroinformatics Vol. 18; no. 6; pp. 1033 - 1054
Main Authors Ahani, Ali, Mousavi Nadoushani, S. Saeid
Format Journal Article
LanguageEnglish
Published London IWA Publishing 01.11.2016
Subjects
Algorithms
Annual precipitation
Catchment area
Catchments
Cluster analysis
Clustering
Drainage area
Estimates
Feature maps
Flood frequency
Frequency analysis
Frequency distribution
Fuzzy sets
Hydrology
Objective function
Precipitation
Regions
River discharge
Runoff
Runoff coefficient
Self organizing maps
Validity
Vectors
Indiana
United States > US
Online AccessGet full text

Cover

More Information
Summary:Cluster analysis methods are a type of well-known technique for regionalisation of catchments to perform regional flood frequency analysis. In this study, a fuzzy extension of hybrid clustering algorithms is evaluated. Self-organizing feature maps and four hierarchical clustering algorithms were used to provide the initial cluster centres for fuzzy c-means (FCM) algorithm. The hybrid approach was used for regionalisation of catchments in Sefidroud basin based on feature vectors including five catchment attributes: longitude and latitude, drainage area, runoff coefficient and mean annual precipitation. The results showed that according to the values of both the objective function and the cluster validity indices, the performances of FCM algorithm often was improved by using the proposed hybrid approach. Also, it was evident from the results that in the case of minimizing the objective function, the combination of Ward's algorithm and FCM provided best results, but according to the cluster validity indices, other hybrid algorithms such as combinations of single linkage or complete linkage and FCM algorithm presented the most desirable results. In addition, according to the results, there are two well-defined homogeneous regions in Sefidroud basin identified by all the examined hybrid algorithms.
ISSN:1464-7141
1465-1734
DOI:10.2166/hydro.2016.239