Geek asks and answers

Recommender Algorithm

I need to make a recommendation algorithm for a website. I came up with a very simple way to achieve this, but I wondered if anyone could point me to some literature or something that could help me better understand how other examples were compiled.

I was told about features like collaborative filtering, clustering, and categorization in Apache Mahout, but I don't know how Machine Learning fits into all of this. I can see how to make algorithms for the above (except for Machine Learning), but wondered if anyone knew about anything else that could be added to the mix.

In addition, what would you say, the purpose of the Recommender, how can it function best? Anyone who wants to share a definition?

Thank!

+5
source share
3 answers

There is an article that discusses the various possibilities of combining different algorithms and creating recommendations. The authors analyzed 37 different systems and their links and sorted them into a list of 8 main dimensions.

Despite the fact that the document was published in 2003, and some of its examples are not available now, it can still be a very good starting point for researchers to create their own recommendation system.

I would like to talk about the definition of recommendation systems by Robin Burke recommendations in my article :

, , .

+5

- ( , Data Mining), . , , , ..

, (, ), .

Netflix Amazon. ", , ", , .

- , . , , , . K.

: , . Amazon.com: .

. : . User-Adapted Interaction.

+3

Coursera , Joeseph Konstan of U Minn, . . , , :

 - Rating Systems
 - Content Based Filters
 - Collaborative Systems (User-user and Item-item)
 - Dimensionality Reduction (SVD, its meaning, and how to compute it)
 - Hybrid Systems

SVD ML, , , - .

, Lenskit ( ) . , , , .

+3

More articles:


All Articles