Summary
In this chapter, you have seen the basics of regression modeling. You have learned that there are some similarities between classification and anomaly detection models, but that there are also some fundamental differences.
The main difference in regression is that the target variables are numeric, whereas they are categorical in classification. This introduces a difference in metrics, but also in the model definition and the way the models work deep down.
You have seen several traditional, offline regression models and their adaptation to working in an online training manner. You have also seen some online regression models that are made specifically for online training and streaming.
As in the previous chapters, you have seen how to implement a modeling benchmark using a train-test set. The field of ML does not stop evolving, and newer and better models are published regularly. This introduces the need for practitioners to be solid in their skills to evaluate models...