Nowadays entertainment is a medium where people can take a break from their busy lives. And movies are one of the sources of entertainment, but the problem is finding your desired content from the ever-increasing millions of content every year. However, recommendation systems come much handier in these situations.
Recommendation systems are predicting systems that radically recommend items to users or users to the items, and sometimes users to users too. Tech giants like Youtube, Amazon Prime, Netflix use similar methods to recommend video content according to their desired interest. As the internet contains huge loads of data, finding your content is very difficult and can be very time consuming, thus the recommendation plays an important role in minimizing our effort.
These systems are getting more popular nowadays in various areas such as in books, videos, music, movies, and other social network sites where the recommendation is used to filter out the information. It is a tool that is using the user’s information to improve the suggestion result and give out the most preferred choice. User/Customer satisfaction is key for building the tool. It is beneficial for both customers and companies, as the more satisfied the customer is, the more likely he/she would want to use the system for their ease, which would ultimately make revenues for the companies. Recommendation system should always be improved as the user choice can differ from other users and if the user is not happy with the result, he/she might not use it again which is the case we want with our system.
Although there are a lot of algorithms, collaborative filtering is the most popular ones used by the companies as it involves user’s interactions more. Collaborative filtering can predict better than content-based filtering because it analyses the user’s browsing history and compares with other users and then suggests results. Whereas the content-based filtering takes the user’s information as an input and finds similar movies and recommends them in descending order using cosine similarity. There’s another method named context-based filtering where it extracts more information from the user like mood, release date, genre, etc to give more efficient results. Our goal in this project was to keep our system very accurate compared to other recommendation techniques while making it as simple as possible. Both the Content-based filtering and collaborative filtering has some drawbacks and lack in accuracy and preciseness. So the proposed system is the Hybrid recommendation system which combines both of the results to give an accurate choice.