"Clustering files"

hun

Hello!

I have a question about clutering. I have some GPS data (the most important are longitude and latitude attribute). I would like to cluster the files, but not in rows.

E.g.: I have 10 csv files (contains lon and lat attributes with 1000-2000 rows). 8 routes are similar (home -> work) and 2 routes are different (home -> shop and work -> restaurant). I would like that: cluster0: the 8 similar routes, and cluster1: the others. Or cluster0: 8 similar, cluster1: other one, cluster2: other one.

Any idea, how can I do that?

Thanks in advance.

Adam

MartinLiebig

Hi Adam,

i think you can simply calculate statistics per file using loop files and aggregate. Stuff like
#Entries
#Most common entry
...
And then use clustering on those 8 examples. At least that would be my first guess.

~Martin