Re: Looking for Linear Stretch Constant for 1D Function

>> g(k*t) - k*g'(0)*t - g(0) = f(t) - f'(0)*t - f(0)
>> because
>> int(0 -> t, f'(k*t)*dt) = (f(k*t) - f(0))/k
>
> That looks right. The problem, I have just noticed, is that k is the
> unknown quantity. How can we get g(k*t) if we do not know k? With a
> linear system
> g(k*t) - k*g'(0)*t=g(0)+ f(t) - f'(0)*t - f(0)
> (with one equation for each t), there would be N equations and N+1
> unknowns.

you'll have to interpolate a function through g(t) so you can predict the
value g(k*t) for a given t
you also need to do that to estimate g'(0) and f'(0)
one equation is enough, but you can use all N to minimise the error due to
the noise and the errors in the estimation


.

Relevant Pages