Hi, I am building a social networking site and looking to implement amazon like recommendation engine around user profiles. What algorithm and method would you people recommend for using? I am using Python for building my application.
The secret behind Amazon.com's recommendation system is a large information data that they have about user shopping. They know from what you have bought what other people also bought. So, based on this behavior they show what similar books you are likely to buy.
So, the first attempt at building a recommendation engine would be to first start from the similar content which would come by comparing the text in both the profile, books whatever you intend to recommend. Compare with what? What you current are? This can come from the information you have filled during registration, links you have clicked or views, things you have recommended or bought. Got it?
Another intriguing idea, I believe carries the potential could be to explore delicious tags and find keywords/similarity based on that relation to come up with a sound recommendation engine.
1 year 51 weeks ago
The secret behind Amazon.com's recommendation system is a large information data that they have about user shopping. They know from what you have bought what other people also bought. So, based on this behavior they show what similar books you are likely to buy.
So, the first attempt at building a recommendation engine would be to first start from the similar content which would come by comparing the text in both the profile, books whatever you intend to recommend. Compare with what? What you current are? This can come from the information you have filled during registration, links you have clicked or views, things you have recommended or bought. Got it?
Another intriguing idea, I believe carries the potential could be to explore delicious tags and find keywords/similarity based on that relation to come up with a sound recommendation engine.