Question

Lets say you have a 2D bounding box with certain dimensions. How could you weigh the data within the box such that data closer to the edges would have less affect on the mean of the data and data towards the center would have a higher weight on the mean? Thank you.

Answer 1

This can be done by superimposing the given data over the 2D normal distribution curve, by doing this the weight assigned to the center most element will be the maximum and it then keeps on decreasing as we move away from the center radially. Thus the data closer to the edges will be having less effect on the mean of the data and the data belonging to the center will have effect, the amount of effect that you want can be adjusted by changing the mean and standard deviation of the normal curve. Small standard deviation reduces effect of data near edges.