When you visit Amazon, Netflix, or your other favorite websites, or use any modern app such as Spotify or Pandora, you will have noticed that they recommend different items to you. These recommendations are created using recommendation system algorithms. Before machine learning based recommendations systems, the recommendations were generated with rule-based systems. However, with the advent of machine learning and neural networks, recommendations have become more accurate.
In this chapter, we'll learn about recommendation systems. We'll use the Retailrocket dataset to implement a recommendation system in two different ways, using TensorFlow and Keras.
The following topics will be covered in this chapter:
- Recommendation systems
- Content-based filtering
- Collaborative filtering
- Hybrid systems
- Matrix factorization
- Introducing the Retailrocket dataset
- Exploring the Retailrocket dataset
- Preprocessing the data
- The matrix factorization model...