Re: Surface Smoothing

*To*: mathgroup at smc.vnet.net*Subject*: [mg127638] Re: Surface Smoothing*From*: "Nicholas Kormanik" <nkormanik at gmail.com>*Date*: Thu, 9 Aug 2012 03:53:03 -0400 (EDT)*Delivered-to*: l-mathgroup@mail-archive0.wolfram.com*Delivered-to*: l-mathgroup@wolfram.com*Delivered-to*: mathgroup-newout@smc.vnet.net*Delivered-to*: mathgroup-newsend@smc.vnet.net*References*: <jvo0fu$cfq$1@smc.vnet.net> <5021312C.1030703@KevinMcCann.com>*Reply-to*: <nkormanik at gmail.com>

Good point, Kevin. The spiky behavior does make things rather scary. By incorporating additional factors (beyond the two in the contour map) I hope to minimize the bad spikes, and maximize the good ones. For now, though, I hope to find a relatively decent "neighborhood," should such actually exist - i.e., the "sweet spot." An analogy might be: A dangerous minefield. If I absolutely have to walk through it, I'd like to try to ascertain the path with the lowest probability of being blown up. Nicholas Kormanik -----Original Message----- From: Kevin J. McCann [mailto:kjm at KevinMcCann.com] Sent: Tuesday, August 07, 2012 9:16 AM To: nkormanik at gmail.com Subject: [mg127638] Re: Surface Smoothing Any smoothing implicitly assumes that you "know" what the data should look like. So, I assume that you know that the spiky behavior is not "correct". Given that, how about a LSQ fit to some satisfactorily smooth function, e.g. a 2d polynomial or a truncated Fourier series? Kevin