The previous two chapters showed how you how to build, test, and optimize recommender systems using R. Although the chapters were full of examples, they were based on datasets provided by an R package. The data was structured using redyal and was ready to be processed. However, in real life, the data preparation is an important, time-consuming, and tough step.
Another limitation of the previous examples is that they are based on the ratings only. In most of the situations, there are other data sources such as item descriptions and user profiles. A good solution comes from a combination of all the relevant information.
This chapter shows a practical example in which we will build and optimize a recommender system, starting from raw data. This chapter will cover the following topics:
Preparing the data to build a recommendation engine
Exploring the data through visualization techniques
Choosing and building a recommendation model
Optimizing...