Abstract: With the increasing number of candidate services that meet the same function on the Internet, service selection becomes more and more difficult, and service recommendation becomes the key issue that needs to be solved urgently. However, the traditional service QoS prediction based recommendation method pays less attention to the role of the service ranking to the recommendation list, which can not accurately reflect the user preference. To solve the above problems, this paper proposes a QoS ranking learning based service recommendation algorithm. It selects low computational complexity listwise loss function to optimize the matrix factorization model, and further improves the accuracy of QoS ranking by mining the neighbor information between users. Experiments on real datasets show that the proposed algorithm has good performance.

Key words: computer application service recommendation, collaborative filtering, learning to rank, matrix factorization