5-second situation-specific recommendation engine for movies
According to research, people spend 18 minutes on average to pick a movie to watch on Netflix. For students and career-focused individuals who do not have much free time, this is far too much time spent on picking a movie, and not enough time spent enjoying the movie.
DECIDR uses a slightly altered K-Nearest Neighbours machine learning algorithm to determine clusters of users who share similar tastes in movies, and clusers of similarily rated items based on a number of factors. Using this information it generates a list of 10 movies that the user should be interested in. With more users and more ratings over time, the predictions will become more accurate. In order to train a model to make predictions with, we used a sample dataset based on approximately 100,000 movie ratings by over 1000 different users. Our sample data comes from GroupLens.