So this is the movielens homepage at movielens.org, where you can sign up or sign in. I will sign in. And since I've already signed in, it takes me straight to the main page, where I can see a few different things. Right away, we can see that I have some recommendations. Top picks. And these movies are movies that Movielens thinks that I'm going to like to watch. It also shows me for each one a predicted rating. How much it thinks that I'm going to like that movie on a scale from one to five stars. I can also then go, and I can rate a movie. Just by clicking on the number of stars that I want to rate it. So, and it will save the rating and use that in my future recommendations. We also see a few other recommendation lines here such as recent releases which is just what's been come out currently. A list of movies that it thinks would be useful for me to rate in order to improve my recommendations. As well as favorites from the last year, which is movies from the past year that Movielens users have rated highly. And then finally my wish list which is movies that I've marked that I would like to see some day. So, there's various different lines, some of which are intended for recommendations, some of which are intended for the user's own reference as they interact with their movies. All of the recommendations are produced based on these five star ratings with some additional data, such as popularity, how many people have rated the movie, etc. Movielens also allows you to exercise some control over the recommendations you get. Up here I can see some basic statistics about my ratings such as I have rated 322 movies. And it also allows me to pick which recommender I want to use. Right now I'm using a recommender called The Wizard. I can also pick The Warrior. In which case it tells me that I'm changing recommenders, and shows me a different set of recommendations. Very different set. I can also pick non personalized recommender called the peasant, that does not take my past preferences into account. And there's additional recommendation options as well. So, In addition to picking the recommender that I want to use, I can configure the recommender to tweak aspects of its recommendation. Particularly, I can say, well, okay, these recommendations are basically good, but I want less popular movies. And so it reorders my recommendations to take out some of the popular movies and add in more movies that aren't quite as popular. And as we can see it changes which movies are picked up at the top picks. So Movielens allows us to use different kinds of customization. So that the user can have a fair amount of control over the recommendation. Also, when we go into a movie's details page, we can see again the predicted raring Displayed both numerically and as stars. And we can see more detailed average statistics, the average rating, and then, the number of ratings that that average is computed from. As well as information about the movie, such as its run time, its rating, and its synopsis, etc. Movielens users also tag movies so that to describe them with various characteristics such as dark humor or anti-hero. And then using tags we can see similar movies. So, these are movies that are similar as far as how users have tagged them, the words that have been used to describe them. We can browse by these similar movies and then we can refine our browsing by saying maybe I want like Deadpool but less violence. And it will show me a variety of typical superhero blockbusters. And I can add in more things, like okay, less violence, but more stylized, and it will continue to refine the kinds of browsing I can do. And it winds up saying The Incredibles is like Deadpool, but less violence and more stylized, which seems somewhat reasonable to me. So, those are a lot of the features that we have available in Movie Lens. The recommender is integrated throughout. So, if I search for Batman, it's going to show me a number of Batman movies but it shows me in here, it continues to show me, everywhere, what I've rated it in orange or my predicted rating in blue. It also allows me to change how I sort the search results by recommendation, number of ratings, when it was released, when it was added to MovieLens etc. So, MovieLens allows users to find movies that they want. They provide ratings, it also serves as a log of the movies they've seen and how much they liked them. They can go back and they can review statistics about the ratings that they've seen to see whether they like common movies or or rarer movies. The distribution of their ratings. As well as draws the history of the ratings they give and the movies that they're rating, the genres they like, etc. To advise users with a way to do fairly detailed exploration of movies, to track their movie watching and get recommendations and also contribute to science since it's an ongoing research platform. With that, thank you and we hope you enjoy the rest of this class.
