Inspiration:
We obtained inspiration from Netflix, Youtube, and many other online movie and video streaming services that give you video recommendations based on your previous searches.
What it does:
Our project takes in a user input for a movie title (restriction: has to be within our dataset of 5000+ movies) and immediately receives a definite output set of movie recommendations based on the one they watched or like to watch.
How we built it:
We used several machine learning tools such as sklearn with its cosine_similarity and CountVectorizer functions (to detect similarities between words) as well as the pandas dataframe to store data read from our movies' CSV file.
In essence, by looking at how similar the movie's plot keywords, genres, and directors are, we can determine which movies are best suited for the user based on how similar these attributes are to the original movie's attributes.
Challenges we ran into:
Learning the process of "vectorizing words" and also understanding how the CountVectorizer() works.
Accomplishments that we're proud of:
- Being able to work through the challenges we faced and not giving up when our other ideas seemed to not work.
- Accomplishing so much for our first hackathon.
What we learned:
We learnt about the general process of machine learning w/ computational linguistics and how it can be applied to accomplish and solve a variety of problems.
What's next for an idea:
Making this work with an even larger database of movies, T.V shows, or possibly music.
Log in or sign up for Devpost to join the conversation.