Re: fixed points piecewise-linear fitting

[Sergei Steshenko]

----- Original Message -----
> From: Macy <address@hidden>
> To: address@hidden
> Cc: 
> Sent: Saturday, March 17, 2012 8:50 PM
> Subject: Re: fixed points piecewise-linear fitting
> 
[snip]
> 
> You might try something, look at the Fourier Transform of the resulting 
> solution.
[snip]

Folks,

let's stick to the KISS  principle.

I have a solution which works quite well for me. I.e. in practical terms the 
whole issue is not burning at all.

I thought somebody could be _practically_ familiar with already invented wheels 
in the field and could point me to articles with "fixed points piecewise-linear 
fitting". I mean, to as simple as possible articles. And/or to precooked code 
with explanation which could be understood by an electronic engineer - not by a 
statistician.

I did go through some articles I found on the web - I don't remember Fourier 
transform or convolution mentioned in them.

I think the issue of derivatives is irrelevant here. I.e. the task is to fit 
analytically unknown function (so nothing is known about its derivatives) with 
piecewise-linear function which by construction has step-like (or bar-like ?) 
first derivative.

Thanks,
  Sergei.