Page 1 of 1

Bug in Crqad.m

Posted: Fri May 13, 2016 11:25
by Lopez
Hi,

First of all, thank for sharing this Toolbox with us.

Im using crqa to work with categorical time series. However, when i was using the crqad tool i wasn't getting the same results i obtained with the R toolbox. So i decided to follow the code and i found the following line:

x=crp2(x1,x2,m,t,e,method,'sil','nor');

When using categorical time series we are not interested in normalising the data and it seems it was doing it anyway. I think if i am right this 'nor' have to be changed to normalisation, which gets the 'non' or 'nor' depending on the user interests.

I hope this helps.

Best Regards,

David

Re: Bug in Crqad.m

Posted: Wed Jul 13, 2016 08:14
by Norbert
Hi David,

thanks for the bug report. Yes, you are completely right. This code was not used so much, therefore I missed to include this "normalization" option in this function (although in the help text it is stated that this would work). I have now fixed this problem (R30.5). crqad allows to select "normalization" or "nonormalization".

Best
Norbert