RECURRENCE PLOTS

Hi,

While I am working on Recurrence Network for 2 months, I am curious how to classify or recognize two time series signals. In many papers, they used Machine Learning or Wavelet Transform. Is it possible that Recurrence can classify the class of time series data?

Thanks for helping,
Tony

Which classes you would like to distinguish? The visualisation of the recurrence plot already exhibits typical features of some classes. However, some classes might not be distinguishable. Even the quantification might sometimes fail.

But it would be good to look at certain examples. Do you have some in your mind?

I am interested in EMG signals. I found some papers which they showed how to classify EMG signals such as Classification of EMG signals using wavelet neural network (2006). Finally I found many papers which they used RQA to classify time series but I have some questions about RQA and Recurrence Network.

I read it on your paper. Is it possible that we compare some parameters from RQA to Recurrence Networks?

1. The mean diagonal line length <L>
2. Determinism <DET>
3. Entropy <ENT>

Since you transform RP to Network, how could you identify these parameters in Recurrence Networks?

Thanks for your helps,
Tony

Dear Tony,

if I understand correctly, you mean how to apply the RQA measures on a recurrence network? This is not necessary. The recurrence network is represented by the recurrence matrix. If you apply the RQA measures, then there is not difference between recurrence plot or recurrence network.

However, you should be aware that usually in a network you can reorder the nodes, i.e., in our case the time points. Such reordering would change the RQA (but not the network measures).

OK .. I got it. I will try to investigate on Recurrence Networks again. My goal is to analyze time series using recurrence networks. I hope you don't mind to answer my stupid questions. Thanks again

No problem. You are everytime welcome with your questions.