Weights for TCEP dataset

[ArnoVel]

Hi, Many different papers related to bivariate causal discovery discuss the necessity of attaching a weight to each pair to account for the fact they come from the same joint distribution.

I do not see this as an option currently in CDT.

Would this possibly be an option in later releases? :)

Thanks!

[ArnoVel]

As a reference:

0.166,
0.166,
0.167,
0.166,
0.143,
0.143,
0.143,
0.143,
0.143,
0.143,
0.142,
0.5,
0.25,
0.25,
0.25,
0.25,
0.5,
1,
1,
0.166,
0.167,
0.333,
0.333,
0.334,
0.125,
0.125,
0.125,
0.125,
0.125,
0.125,
0.125,
0.125,
0.2,
0.2,
0.2,
0.2,
0.2,
0.25,
0.25,
0.25,
0.25,
0.5,
0.25,
0.25,
0.25,
0.25,
1,
1,
0.333,
0.333,
0.334,
0,
0,
0,
0,
0.083,
0.083,
0.084,
0.083,
0.083,
0.084,
0.083,
0.083,
0.084,
0.333,
0.333,
0.334,
1,
1,
1,
0,
1 ,
0.083,
0.083,
0.084,
1,
0.5,
0.3333,
0.3333,
0.3334,
0.3333,
0.3333,
0.3334,
1,
1,
1,
1,
1,
0.25,
0.25,
0.25,
0.25,
1,
0.3333,
0.3333,
0.3333,
0.2,
0.2,
1.0,
1.0,
0.5,
0.2,
0.2,
0.2,
0 0.5,
1
1
1

is the list of weights for the current dataset (108 pairs). The current TCEP version differs from the CDT one in the following:

• pairs 52 53 54 55 missing (not all of them are multivariate, strange) all indexes after 51 are offset by 4
• pair 71 missing indexes after 71 offset by 5
• last pair in CDT is 104. 104-5 gets us the 99th pair.

[ArnoVel]

edit: Most missing pair have a corresponding weight of 0.

[diviyank]

Hi, The difference between the datasets comes from the version of the Tuebingen Cause-effect-pairs datasets. I might update that as well. I will update the weights very soon, thanks for the contribution !