UnboundLocalError: local variable 'pvalues' referenced before assignment

[ghost]

Hi, I am trying out factor models and ran into this error:

UnboundLocalError: local variable 'pvalues' referenced before assignment

which is coming from this line

Riskfolio-Lib/riskfolio/ParamsEstimation.py

Line 204 in 4087c5d

value = pvalues[pvalues.index != "const"].max()

Probably because for my dataset I didn't run into this condition:

Riskfolio-Lib/riskfolio/ParamsEstimation.py

Line 199 in 4087c5d

if best_pvalue > new_pvalues[i] and cond_1 <= threshold:

maybe an else condition could help to handle it.

Also, is it uncommon for this to occur? Is it a big deal? If so, how do I change my data/factors to ensure this doesn't occur?

[dcajasn]

Hi, I remember that when I was writing that part I left that code on purpouse because I wanted that an error appears when there is no significant variables among risk factors. Meanwhile I check that part, you should try to add it more risk factors or change the criterion to AIC or BIS.

[ghost]

will try it out thanks