It is from Kaggle Competitions where the training dataset is very small and the testing dataset is very large and we have to avoid or reduce overfiting by looking for best possible ways to overcome the most popular problem faced in field of predictive analytics.
