Divide a n-dimensional domain into partitions/cells and average points in each cell. Very useful for scattered data interpolation.
This function is useful for data that contains dense data in some location in a domain and not in others. When trying to interpolate n-dimensional data, the desire is for the data be equally space. Often this is not the case for test data. To interpolate in n-dimensional space more accurately, a local average of points can be computed using this function to better facilitate widening the distance between points where the points are close together. This also deals with points that are close together but have very different functional values w = f(u). I found this helps in n-dimensional scattered data interpolation.
Run the example contained in the help of this function to better understand its behavior.
Note that this is similar to n-dimensional smoothing and resampling filter.
- The partition index was not correctly calculated. on line 147.
- uScaled needed adjusting so that the partition index would be correctly calculated on line 112.