Netflix's Personalized Movie Recommendation System
A Netflix's Personalized Movie Recommendation System is an personalized movie recommender system that supports a Netflix service.
- Context:
- It can make use of Netflix Taste Communities Netflix Taste Clusters.
- It can be developed by Netflix Recommender System Engineering Team (which includes an ML engineering team).
- …
- Example(s):
- Counter-Example(s):
- See: Netflix Prize, Content Recommendation System, Item Recommendation System.
References
2018
- http://www.vulture.com/2018/06/how-netflix-swallowed-tv-industry.html
- QUOTE: ... It has replaced demographics with what it calls “taste clusters,” predicating programming decisions on immense amounts of data about true viewing habits, not estimated ones. It has discovered ways to bundle enough niche viewers to make good business out of fare that used to play only to tiny markets. ...
... taste communities are the tool Netflix relies on to drive viewers to new material it estimates they might want to watch. The best explanation of how this works that I heard during my time at Netflix came during a new-showrunner orientation. Whenever new producers start a project, Netflix gathers a dozen or so representatives from its various departments to give them a presentation on how the company works. The Tuesday afternoon I sat in, the team from writer Susannah Grant and producer Katie Couric’s upcoming limited series Unbelievable was getting the rundown. ...
... The Netflix algorithm figures out which taste communities a member is in and then pushes the shows it thinks those members will enjoy to the top of their home screen. “We have a saying: Your Netflix is not my Netflix,” De Carlo says, noting that taste communities aren’t some static construct, either. “Most people are usually members of a few different communities,” she says. “We’re complex beings, we’re in different moods at different times.” ...
... If verticals are the way Netflix executives think about what kinds of content to buy or make, taste clusters help them analyze how subscribers interact with programming. The phrase, along with the interchangeable “taste communities,” comes up time and again during my visits. Instead of grouping members by age or race or even what country they live in, Netflix has tracked viewing habits and identified almost 2,000 microclusters that each Netflix user falls into. While it’s not a direct parallel, taste communities are sort of like Netflix’s version of the demographic ratings used by traditional ad-supported networks, just more evolved. ...
... It shows how one of Netflix’s biggest hits, Black Mirror, plays particularly well in two major taste communities: Cluster 290 and Cluster 56. ...
- QUOTE: ... It has replaced demographics with what it calls “taste clusters,” predicating programming decisions on immense amounts of data about true viewing habits, not estimated ones. It has discovered ways to bundle enough niche viewers to make good business out of fare that used to play only to tiny markets. ...
2017
- "This is how Netflix's top-secret recommendation system works.” In: Wired Magazine.
- QUOTE: … Netflix splits viewers up into more than two thousands taste groups. Which one you’re in dictates the recommendations you get …
… More than 80 per cent of the TV shows people watch on Netflix are discovered through the platform’s recommendation system. ...
- QUOTE: … Netflix splits viewers up into more than two thousands taste groups. Which one you’re in dictates the recommendations you get …
2016
- (Gomez-Uribe & Hunt, 2016) ⇒ Carlos A. Gomez-Uribe, and Neil Hunt. (2016). “The Netflix Recommender System: Algorithms, Business Value, and Innovation.” In: ACM Transactions on Management Information Systems (TMIS) Journal, 6(4). doi:10.1145/2843948
- QUOTE: This article discusses the various algorithms that make up the Netflix recommender system, and describes its business purpose. ...
2013
- "The Science Behind the Netflix Algorithms That Decide What You’ll Watch Next.” In: Wired Magazine.
- QUOTE: … The company estimates that 75 percent of viewer activity is driven by recommendation. ...