Rounded values

[DemGrg]

Somehow now the values given by the broken function are always rounded vs exactv (tested lm model):

f.e.
contribution
(Intercept) 2300
tenure = 3 -1900
MonthlyCharges = 74.4 350
SeniorCitizen = 0 14
final_prognosis 720
baseline: 0

while: predict(model, data_in_test[analysed_user,]) gives 721.206

[pbiecek]

This is because signif(digits=2) function was used for rounding.
But I agree that it can be misleading, in the latest version both print() and plot() functions have additional arguments digits and rounding_function so user may request a specific behaviour. And by default it's round(digits=3).

[pbiecek]

Here: https://pbiecek.github.io/breakDown/articles/break_lm.html you can find some examples how to use both arguments

[DemGrg]

excellent, thank you....
works great now