Real-Wrold Goal

Recommendation system has become a common topic in different areas, including online shopping websites, video streaming platforms and online news providers. These websites and platforms make use of different models to find out the most attracctive contents for their users, and display these contents to them for making more profits. However, this kind of classic recommendation system has several drawbacks, including limited contents and views, unfair possibility of being recommended and stale recommended contents for users.

In this way, my goal in this project is to come up with a more just and diverse recommendation skeleton for users. In my recommendation system, I would try my best to provide equal chance of different contents being recommended, new contents that could go beyond users' mind and the most-up-to-date contents that could reflect users' most demands. I would make these points clearer in following analysis.