How to get the most out of your models: overcome overfitting!

Use these techniques to get the best performance out of your model. A common pitfall in machine learning modelling is overfitting. Overfitting happens when the model learns the training data too closely, causing it to perform poorly on new data. A common workflow to identify overfitting is to, train the model and record the performance on a test or holdout data. If the difference in the metrics of train and test set differ by a large margin, then it is safe to assume that the model is overfitting.