How to take pairwise Transfer Entropy ?

Humair Ali

unread,

May 17, 2020, 7:08:42 PM5/17/20

to IDTxl

Hello all,

I am new learner of Transfer Entropy, I configure the IDTXL tool, as well as run all demos.

Now I want to do in TE,

Pairwise TE between each pair in R: TE1-2, TE1-3,...., TE1-186 (There are 186 variables in R with 5001 samples each) how can i take Transfer Entropy in python.

also, there is not so much tutorial or codes on the internet regarding Transfer Entropy.

Thanks

Patricia W

unread,

May 18, 2020, 4:41:17 AM5/18/20

to IDTxl

Hi,

I would recommend to get started by adapting the bivariate TE demo script to your data: https://github.com/pwollstadt/IDTxl/blob/master/demos/demo_bivariate_te.py

Bivariate TE calculates TE pairs of sources in the network. If you follow the demo and call analyse_network() on your data all possible pairs of nodes are evaluated using bivariate TE. For more information, you can start with the Wiki and the literature referenced there. You can also always post questions here.

Best,

Patricia

Humair Ali

unread,

May 18, 2020, 9:35:01 AM5/18/20

to IDTxl

Thanks Patricia for your response,

i have one more question about graph/matplot, when i use my data on demo code it will be generated same graph in edges.

how can i change graph/plot ?

Reply all

Reply to author

Forward