Hi,

I am quite new to Data Mining. I am using Machine Learning and Random Forest
to classify percent cover on the ground. I have inserted a bunch of predictors in the process. The questions that I got from my audience is "data mining" generate different models each time as it picks different predictors to find the best fit and will lead to overfitting problem as well. How do I defense it? Or is data mining actually good for modeling or I should only use multiple regression?

Thanks.