I am working on a project 'Rossmann Sales prediction', in which I have to forecast the sales of Rossmann Stores. So it is a supervised ML problem. I applied random forest. But then in interviews question arises like - 1) why you did not apply Linear Regression? 2) Why you directly applied random forest? 3) Why did you not choose boosting technique?
So how to defend this kind of question? Which algorithm works well and what reason to choose that algorithm?