Modified Clustering Mechanism

[PaulPrevedoros1]

Has anyone tried a modified clustering method.

What I am trying to do is to map settlement pattern of archaeological sites. The clustering algorithm use Euclidean distances. However what I want to do is to modify the clustering to use cost distance as the distance between the points.

I suspect this will produce different cluster patterns accounting for terrain, road network, and barriers.

Has anyone done this or have any suggestions, do I have go into the unknown by myself

Cheers     

[DanPatterson]

How Multivariate Clustering works—ArcGIS Pro | Documentation

How Spatially Constrained Multivariate Clustering works—ArcGIS Pro | Documentation

Do you have an attribute that you could throw into the mix to modify the result?  What "cost" are you thinking about?  Is there some directionality in your observations? 

[PaulPrevedoros1]

I have gone the process of doing the QQ process of the of cost distance to all other point (limited search) so I need to create the algorithm to do the next step, I was hoping to avoid this, since I am also .. sort of retired The mathematical brain doesn't work as well as it used to.

... sort of retired the second