You’ll find recommendations in this chapter, but they’re not personalized. That doesn’t mean that this chapter will be less important.
- You’ll learn that using non-personalized recommendations can also show interesting content.
- You’ll see examples that show why your site should order content, and you’ll learn how to build charts to show users what’s popular and that highlight items of interest to other users.
- You’ll learn how to calculate association rules by creating itemsets based on the shopping basket and then use those rules to create seeded recommendations.
- You’ll see how the recommender component is implemented, which is the core component in the MovieGEEKs example site and the one that provides the recommendations.
Non-personalized recommendations are usually where most sites start because it’s easy and doesn’t require that you know anything specific about the users. Non-personalized recommendations are good because you can always show those, despite how little you know about the users. People might say that non-personalized recs should only be shown until the system knows enough about the user to show more personalized recs, but always remember that humans are flock animals by nature, so most will be suckers for knowing what content items are the most popular—if for no other reason than to ensure what not to like.