parameters - Does minpts=4 is the best setting for any dataset using DBSCAN algorithm for clustering? -
the article on dbscan "https://www.aaai.org/papers/kdd/1996/kdd96-037.pdf" explains minpts value must 4 datasets on dbscan being used clustering data points. gives best results eps value??
in later work, authors suggest use minpts = 2 * dim default.
j. sander, m. ester, h.-p. kriegel, , x. xu. 1998.
density-based clustering in spatial databases:
algorithm gdbscan , applications.
data mining , knowledge discovery 2, 2 (1998), 169–194.
http://dx.doi.org/10.1023/a:1009745219419
if have duplicates, use larger value: "our experiments indicate value works databases d each point occurs once, i.e., if d set of points."
smaller values more computationally efficient. thus, keep minpts small not small.
always study result. never use without double checking.
Comments
Post a Comment