Cloud service recommendation based on unstructured textual information

With the rapid growth of cloud service platform, it is more and more difficult for users to discover cloud services satisfying their personalized demands. The unstructured textual information, such as the service descriptive texts and service tags, contain rich features of cloud services and are useful for cloud service platforms to construct personalized service recommendation. This paper proposes a two-stage model for cloud service recommendation by integrating the information of service descriptive texts and service tags. In the first stage, we propose a Hierarchical Dirichlet Processes (HDP) model to cluster cloud services into an optimal number of clusters based on descriptive texts. In the second stage, we propose a Personalized PageRank algorithm based on service tags to rank and recommend cloud services in each cluster. Our experiments on a real data set show that the proposed two-stage model can segment cloud services well and obtain accurate recommendation results compared with the baseline methods.