We introduce you to the big world of recommender systems. In a nutshell, a recommender system is an automated system that filters some entities. These entities could be products, people, ads, movies, or songs. It uses user preference to predict items the system thinks the user will like.
We cover recommender systems in-depth in our Data Science Bootcamp.
What You'll Learn
- What is a recommender system, why they are important, and how they work?
- Why big companies like Amazon, Netflix, Pandora, and YouTube rely on them to serve you the most relevant content?
Hi, and welcome back to data science in minutes. I’m Blaire and in this video, we’ll be learning about recommender systems.
So, what do you think of when you think of recommendations? My friend recommends that I watch “Stranger Things” and my sister recommends that I watch “Ozark” and my mom recommends that I see this new documentary on “Planet Earth”. Are these recommendations based on what they like or what I like? And what if I want to watch something different that none of those three recommended? Wouldn’t it be great if someone just understood me? Well, that’s where recommender systems come in.
So, what really is a recommender system? A recommender system is really an automated system to filter some entities these entities can be any products, ads, people, movies, or songs and we see this from all over on a daily basis from Amazon to Netflix to Pandora to YouTube to eHarmony. For example, we watch a movie and then later on we get a recommendation for a different movie based on the power of previous viewing history it could also be a product that we bought and then we get a recommendation for another product based on the previous product viewing or purchase history and the recommender doesn’t work only in what products we are being shown, but also in what order the products are being ranked.
Why are recommender systems being built? Businesses are showing us recommendations and relevant content for a couple of reasons. For one, most businesses think they understand their customer, but oftentimes customers can behave much differently than you would think so it’s important to show the users what is relevant to them while also sharing new items they would be interested in Recommender systems also serves to help us solve the information overload problem and helps us narrow down the set of choices and for businesses they get the benefit of selling more relevant items to the user It is also there to help you, the customer, discover new and interesting things and to help you save time and from a business perspective it helps to better understand what the user wants.
So, how does it work? Let’s say you like cheeseburgers sans tomatoes. How does your mom know whether or not you like tomatoes on your burger? She could have asked you specifically if you want Tomatoes or she watched you pick them out every time you order a burger. A recommendation engine works the same way. It will either ask you what you want or ask if the content was relevant, look at other users with similar behavior, or study your activity. And even if your mom knows you well maybe a machine learning algorithm knows you better. For example, when we go to Netflix or any other service that relies on recommendations the first time when we go there they will ask you what are your taste preferences and there is a reason for that. Because if they do not know what your taste preferences are at all it’s a “cold start problem” they have no idea and they have no profile for you and they will force you to at least put in something.
What movie do you like and so on because how would they know otherwise? Let’s say I place an order through Amazon Prime for bananas now I’m being shown a banana slicer in the recommendations to me. This banana slicer has over 5,000 reviews in a five-star rating. User ratings, number of reviews, and relevancy can play a factor in terms of what is being recommended to me. Another example of this is how YouTube’s recommendations work. Videos that have a lot of watch time, engagement, is relevant to the topic I’m searching for, and is a relevant topic based on my video history will be shown to me. If I watch a trailer for “Avengers Endgame” I might be shown a trailer for “Iron Man” or funny bloopers from filling the movie or an interview with some of the actors.
These algorithms are so smart that they are able to decide what to show us and it can be scarily accurate. It should also be noted that each company might have their own algorithm and way of generating recommendations and that what one company’s method for applying recommendations does not apply to all. In our next video, we will look at the different types of recommendations and how they are applied. Give us a thumbs up if you like this video, leave us a comment about the strangest thing you’ve been recommended, and what topic you’d like to see covered in less than 5 minutes.
For more tutorials check out online.datasciencedojo.com. Thanks for watching and we’ll see you in our next video.
Blair Heckel - Blair holds a Bachelors degree in Marketing from Washington State University and has a background of leading data-driven marketing campaigns.