Hi,
I'm working with a time series that I have made regular through interpolation of null values where there is no event. There are about 50 000 data points scattered through the 525600 sequence. I seek to do some RQA analysis without stripping out the nulls or using common regularization strategies like substitution with local or global averages. I'm comparing actual data to (sparse) simulated chaotic Henon data.
I'm having problems with VRA crashing out on data input. I'll try with CRP when I can download it. Is there any reason that the programs would reject data with null values?
My reading of the RQA functions is that they should be resistant to nulls in the data since they only act when there is data at the intersection of the different representations of the series. Is that correct or is there opportunity for divide by zero errors?
Regards
Red