Hi all, I'm working on creating a Movie Suggestion AI for a graduate project and wondered if the AI hobbyists/professionals could provide some input.
What I intend to do is have an interface where the user puts in 2 movies and the result will be a movie with common traits to both. Doing this is fairly simple just using keywords and seeing which are the most common and choosing that movie, but I'm struggling with how to turn this into a Machine Learning implementation.
I intend to use Weka to train the algorithm first, then use that model to make decisions for future queries. The problem is, if I simply put in the keywords and tell Weka what the result should be (as training), won't it simply just be learning my method of finding the most common keywords?
That doesn't seem like Machine Learning, can anyone provide a suggestion on how I can train the model so it also will make it's own assumptions/improvements?
[link][5 comments]