There is an answer to this already on here for mullticolline

There is an answer to this already on here for mullticollinearity but ive found many sources online that say heteroscedastic residuals. Help

A log transformation might be appropriate to alleviate which problem(s)?

Autocorrelated residuals

Multicollinearity

Heteroscedastic residuals

Solution

for multicollinearity it is obvious that you can use log transformation but in heteroscedastic residuals also you can use it. it appropriate for both but used widely for multicollinearity.