Neighbor confidence

I just made a movie showing how the weights in an interpolation scheme (i.e., the confidence factor in neighbors) change as a function of the smoothness of the data (that is, as the randomness of the correlation increases: or, as neighbors become less trustworthy).

Neighbor enquiries by position

Here's another one, showing how neighbor weight is a function of the position at which one is interpolating, for a spherical model. Here's the same thing, only for the gaussian model. (They're very similar.)

Less than Abe

I also do a strange kind of map-making in which we throw away some information. This is done because we may have more than is really necessary, or we may have some noise. Watch here as Abe's Face is successively approximated with less information.

Click to continue... to get back to the list of exhibitors...