Personalizied Service Recommendation Based On Context-aware QoS Prediction

Personalized recommendation systems generate personalized recommendations by mining the potential relationship between users and projects and predicting the preferences of users, which has been successfully applied to tourism, music, film and many other fields.However, with the development of cloud computing, various types of Web services based on cloud environment are constantly emerging,which brings new challenges to personalized recommendation technology. It is the main research direction in this paper that how to recommend high quality Web services to meet the needs of users in the same service with the same or similar functions.This paper mainly focuses on personalized Web service recommendation based on context-aware QoS prediction. The main contributions are as follows:Firstly, this paper proposes a weighted context similarity model to model the influence degree of each attribute of the context information to the scene similarity when services being called,according to the binary relation between user and Web service and the context information of Web service when being called as well.Then this paper applies the weighted context similarity to heuristic K-modes clustering algorithm and hierarchical clustering algorithm to find users and services with similar context. Besides, this paper combines user clustering and web service clustering to solve the problem of the sparseness of the user-Web service matrix, and generates web services recommendation via QoS prediction.Finally, in order to solve the problem that the similarity degree model can not accurately calculate the similarity under the high-sparsity condition and can not measure the influence of different context factors to prediction value, this paper introduces factorization machine to directly model the user-Web service binary relation and context information, train the model and predict QoS. Web services recommendation is generated via QoS prediction.In this paper, we use WSdream dataset2, which is a popular dataset of web service recommendation, to predict the response time and throughput. Experimental results show that the algorithm proposed in this paper has a good effect on the two indexes of RMSE and MAE, which proves the effectiveness of the algorithm proposed in this paper.