# linear regression with a transformed dependent variable

Can I use log function with base 10 instead of the natural log in order to transform a dependent variable in a linear regression? By using log we can fix the violation of the normality assumption.

Changing the base of the logarithm is equivalent to a factor in the outcome. Specifically in your case $$\log_{10}(y) = \frac{1}{\ln(10)} \ln(y).$$ It therefore doesn't make a difference for linear regression, except that the regression slope(s) will change by the same factor.