Our social networks needs a fabulous recommender system script written. Our site is written in PHP and use SQL.
Schema as following:
A social network contains many Users and Items. The network will keep track of how much a User checks into each Item and translate that into a score between each Item and User.
For example, User A likes Movie A with a score of (80), Movie B(50), Movie C (45), Sports D (30), Actor E (40), Actor F (30), Actor G (38), Music H (14).
The Items each have their own tags. Eg, Britney Spears (Female singer, Pop singer) The Shinning (honor film) etc. This allows Items being categorized and associated nicely.
Users have friends who each have Items they like.
We need an algorithm that would do a great job recommending accurate Items to Users. As far as we know, there are User/Item based recommendation systems. We would like a system that has a few factors which allows a level of flexibility for us to allocate weight to each factor. Some suggested factors can be 1. User’s own preference for Items 2. Other similar Users’ preference 3. Friends’ preferences 4. Similarity of Items
Collaborative filtering of different users’ behaviour, preferences, and ratings, Automatic content analysis and extraction of common patterns and Social recommendations based on personal choices from other people can all be considered.
If you have experience in recommender system, please share with us. We can provide you with resources needed.
Check out other projects in our profile. Winner will definitely get more work from us:)
[url removed, login to view]
Thank you and have a wonderful day!