Nikhil Ketkar

Earlier I was using KMeans to get discrete locations from continuous 3D points. Dropped the idea as this is going to bias the results. Basically, if I have used KMeans instead of a uniform grid and then I do a test on an unseen sample, the sample is not really unseen as the entire dataset was used to do the discretization.

Here are the points in each location.

Here are the walks going through each location.

Note that in both cased K was set to 1000.

Here are some points from the UCT Transfer Learning dataset in 3D. Its funny that we can almost see the map.

I have been superimposing a grid to get discrete locations from continuous 3D points. The following diagram should give a general idea.

Now the question is how main points lie in each discrete location, and perhaps more importantly, how many walks pass through each location?
Here is a plot for the number of points in each location.

Here is the plot for number of walks in each location.