More digits in tree visualization prediction possible?

[Julia0207]

Hi,

we work with the tree visualization of causalml. We train the uplift tree by UpliftTreeClassifier(), fit it using uplift_model.fit() and visualize the uplift tree by uplift_tree_plot().
Is it possible to have more digits in the mean prediction of the treatment group (control group) within the visualization ?

Thanks and kind-regards,
Julia

[paullo0106]

Thanks for reaching out, Julia. Did you mean to increase the second decimal place for mean treatment and control displayed on the leaf nodes?

[Julia0207]

Yes, that is what I mean. :-)

[paullo0106]

@Julia0207

The decimal point can be changed in uplift_tree_plot() function here, for example, lsY.append('%s:%.3f' % (szX, n)) would give us numbers to 3 decimal places, as shown in the attached. We can add an optional parameter in the function to address different user preferences.

Out of curiosity, Are you working on something that requires high precision of uplift calculations?