I want to perform a linear regression on house price dataset to rank the features that most impacting the price. I handled categorical data by one hot encoding. Then, I transformed (log) all the features except the categorical data (0,1). After I run the regression, is the coefficient of the categorical data is the same scale as other feature's coefficient?
I have a problem to understand why 0,1 data cannot be transformed and is the coefficient of untransformed one hot encoded data is the same scale with other transformed coefficient.
