Well no. The problem is not a recommendation system. The problem is that the streaming companies play keep-away with content. They make you search for it. They stumble through other offerings on your way to the thing you searched for. Unlike a library, you can't just roll up to the card catalogue and browse. Because at no time will they give you a complete inventory of what is available. And there is no guarantee that something that was available before is still available now. Part of the issue is inherent with a on-screen interface. TVs and tablets don't have a decent keyboard interface. And their UI is optimized for playing video, not for complex navigation or filing systems. But it really would help if the services provided the services of a curator. Somebody with some film background, or a subject matter expert for documentaries, who assembles actually programming. A channel you can tune into based on your mood, and possibly be exposed to content you would ordinarily never have found.
I think the issue with this is that generally a very small percentage of users comprise 90%+ of the inputted comments. That would unfortunately bias the data, and in the field of data science you'd end up removing it from any recommendation algorithm.
Well no. The problem is not a recommendation system. The problem is that the streaming companies play keep-away with content. They make you search for it. They stumble through other offerings on your way to the thing you searched for. Unlike a library, you can't just roll up to the card catalogue and browse. Because at no time will they give you a complete inventory of what is available. And there is no guarantee that something that was available before is still available now.
Part of the issue is inherent with a on-screen interface. TVs and tablets don't have a decent keyboard interface. And their UI is optimized for playing video, not for complex navigation or filing systems. But it really would help if the services provided the services of a curator. Somebody with some film background, or a subject matter expert for documentaries, who assembles actually programming. A channel you can tune into based on your mood, and possibly be exposed to content you would ordinarily never have found.
Maybe if they had comment sections this problem would sort itself out.
I think the issue with this is that generally a very small percentage of users comprise 90%+ of the inputted comments. That would unfortunately bias the data, and in the field of data science you'd end up removing it from any recommendation algorithm.
With an upvote and downvote system