Analyses for the paper
Analyses for the paperΒΆ
you will find the following sections:
UCI Performance all anlyses surrounding the data from UCI repository
Reasons for Conofund-Leakage
a) with continuous features
b) low precision features (including categorical)
c) adding more features [added in revision]
Real World Performance
a) main analysis
b) permutation analysis [added in revision]
c) feature importance analysis [added in revision]
d) CR only on healthy [added in revision]
Walk Through Analyses
a) using a binary TaCo
b) using a simulated continuous (causal) confound