2
$\begingroup$

I am trying to calculate the three-term connection coefficients

$$ Λ_{l,m}^{d_1,d_2,d_3} = \int_{-\infty}^\infty \varphi^{(d_1)}(x) \varphi^{(d_2)}_l(x) \varphi^{(d_3)}_m(x) dx $$

for Daubechies wavelets numerically using Python. This report (p. 20) gives a nice overview of how to calculate them.

I wrote a (well documented) implementation in python and for the two-term connection coefficients had no problems but when I compare my 3-term results with the ones of Resnikoff and Wells (1998) I find lots of errors (only the $l=m$ coefficients appear to be correct).

I am starring at the code for weeks now, but I cannot find any errors. I would be very pleased if somebody would give me a hint where the error might be.

Improve this question
$\endgroup$
2
  • 1
    $\begingroup$ you will probably be better off asking at either StackOverflow or SciComp.stackexchange.com $\endgroup$
    – Willie Wong
    Commented Jul 13, 2016 at 17:32
  • 1
    $\begingroup$ I guess StackOverflow is really the wrong place, but I did not know about SciComp! I will give it a try. Thanks! $\endgroup$
    – J. Doe
    Commented Jul 13, 2016 at 19:21

1 Answer 1

Reset to default
0
$\begingroup$

Since I also need to calculate connection coefficients with python I've tried some stuff to get your code working. I have put this script on my github (https://github.com/FalkoGiep/CC_Wavelets).

I have not kept track of what is wrong with your code, but the main difference are that I use scipy.linalg.lstsq instead of the numpy variant. Also I use a different expression for the moments.

Improve this answer
$\endgroup$

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .