Training error< validation error means high variance
Hello Shivani,
Thanks for pointing out. We've made the change. It is correct that training error lower than validation error indicates overfitting and high variance. Even a small change in the in-sample data will have to lead to remodelling as our current model is too dependent on the in-sample data -it's overfitted. That explains the variance in the models we will need to explain even small changes in the in-sample data.