Machine Learning: The Bigger Picture, Part II

This is the second part of Tamis van der Laan’s article featured in the new DZone Guide to Big Data Processing, Volume III. Get your free copy for more insightful articles, industry statistics, and more.  Overfitting So far we have assumed we only have a machine learning model, a training set of samples, and a optimization algorithm to learn from these examples. The next thing we will talk about is the problem of overfitting. If we take our example of a discriminative classifier, we see that it splits the space into two distinct regions for each class. We can also consider a classifier that splits space into multiple regions for classification. Given our example, we get the result in the figure below: We see that the model now splits space into multiple regions and manages to classify more samples…


Link to Full Article: Machine Learning: The Bigger Picture, Part II