Online streaming services like 'Amazon Prime', 'Netflix' uses some Recommendation systems to recommend movies to their users based on movie plot, actors, ratings, user preferences.
Build a recommendation system that recommends movies to the users based on the movie plot, director, actors and Genre
The data is available in the below link
https://query.data.world/s/uikepcpffyo2nhig52xxeevdialfl7
There are three major steps in this project:
Extract the keywords from the plot and combine information from other attributes like 'actors','Director' and 'Genre' into a single column
Convert the text data into numerical data. This process is also known as 'Word Embeddings'
Cosine similarity is a metric used to measure the similarity between the two vectors. Mathematically it measures cosine of the angles between two vectors projected in a dimensional space
We will get a similarity matrix which can be used to predict the similar movies based on the plot of the movie searched
Some platforms allows users to search for movies in anonymous mode. We can recommend movies to those users based on their search without having any user information in our database
The major disadvantage of this method is that user lacks the variety in the suggested movies. He recieves suggetions of the movies having similar plot and genre