Dear All,

I want to ask your advice about clustering method. I have rock type data for 22 catchments and the data consists of the area of each rock type in terms of percentage and also in km^2 (please refer to the excel sheet as per attached). By looking into the data, maybe Houheikyou catchment can be classified into the same cluster as Jyouzankei catchment because both catchments are covered up by andesitic rocks, volcanic breccia, and rock tuff although the percentage is different. Like 90.92 % area of Houheikyou is covered up by andesitic rocks whereas area of andesitic rocks for Jyouzankei is only 57.31%. So I am not sure if actually Houheikyou can really be classified as the same group as Jyouzankei or not.

What I have done is try to classify those 22 catchments by picking up a certain rock type i.e. granitic rocks. Not all catchments have granitic rocks but majority catchments has it. Then by using the information of granitic rocks area in km^2, I did the clustering by using the following command:

In=FindClusters[{11, 42, 45, 19, 110, 8, 23, 73, 59, 54, 3, 11, 3, 8, 62, 6, 51, 1}]
Out={{11, 19, 8, 23, 3, 11, 3, 8, 6, 1}, {42, 45, 73, 59, 54, 62, 51}, {110}}

Result show that those catchments can be classified into 3 clusters. However, I am not convinced with the above method. I meant, I just picked up one rock type as the basis of clustering. I am not sure if it can represent the catchment characteristics or not. Hence please kindly advise if there is any objective way for clustering but so far the information which I have are only catchment area, rock type name and its area for each catchment. Thanks for your kind attention.

Best Regards, Intan

Attachments: