This is my Data Science Capstone Project.
It provides streaming service recommendations based on a user's preferences in movies.
The project asks questions about movie preferences and then rates streaming services based on how they accommodate those preferences.
Streaming services are rated by maximizing the following metric:
Service Score = Σ (user score * percentage of total * mean score) for each entry in the category
Where user score is 1 if the user checked the box for the entry in the category and 0 if they did not,
percentage of total is the percentage of the total movies on a service composed of that entry in the category,
and mean score is the mean IMDb and TMDb rating for that entry in the category.
Genre Example: 21% of a service's movies are action movies and they have an average score of 6.8. The user checks the Action box.
The Service Score for action movies is then: 1*.21*6.8 = 1.428. The total Service Score would be the sum of these for each genre.
Large scores mean the service is a good fit for the specific category.
The service scores then go through a sigmoidal transformation with the function:
sigmoid(Service Score) = 1/(1+e^(-1*(Service Score-5)))
These values were chosen because they make a very low score close to 0 and a very high score close to 1.
The goal of the project is to find the streaming service that best fits the user's movie preferences.
It will do this by finding the service with the highest score.
|
|