Although Amazon ML is set to solve classification and regression problems, the service can also be used for other supervised data science problems. In this last section, we looked at two classic problems: Recommender systems and named entity recognition.
- Making recommendations: A recommender system seeks to predict the rating or preference that a user would give to an item. There are several strategies to build recommender systems:
- Collaborative filtering: This involves using the behavioral patterns of similar users to predict a given user's preferences. It's the other people also bought this approach.
- Content-based filtering: This is the strategy where the features of a certain content are used to group similar products or content.
To use Amazon ML for recommendations, you can frame your solution as a content-based recommendation problem. One way to do this is to extract features for your products and users and build a training dataset where the...