What are more advance recommendation system algorithms

Recommendation systems in e-commerce

Streaming services: Recommendation services from Netflix and Spotify

The video streaming service Netflix only integrated a new recommendation system into the platform at the beginning of 2016. The algorithm that plays films and series to suit their personal taste has been revised. The Personalization algorithms from Netflix doesn't consider demographics like age or gender. The indicator used is the specially collected data. As soon as the account is set up, the user is asked to provide Favorite films and series on. Central questions are answered in the course of usage: What has the customer viewed before? And how did he rate what he saw? By comparing all customers based on their preferences and ratings, the platform then speaks precise recommendations out.

Previously there were problems when the service started in a new country. Because there was no database on which recommendations could have been calculated. The new algorithm therefore works with it cross-border customer groups. Country and region-specific tendencies are still included.

Also the music streaming service Spotify has been working with personal recommendations for a long time. Every week, the service compiles a list of songs that potentially match the user's taste. Of course, “Your Mix of the Week” is also automatically created by algorithms.

Starting points are on the one hand self-generated playlists of other users, on the other hand a precise taste profile that Spotify creates for each user. The service works here with extremely narrow genre definitions. In addition, it uses its own software that analyzes articles and texts on blogs and magazines in order to be able to classify artists as precisely as possible. The recommendation service also recognizes so-called genre outliers that do not fit the overall profile and that a user z. B. alluded to on a whim. Spotify does not take these songs into account in the creation of personalized playlists.