Recommendation Of Cloud Services Based On Markov Process And Multi-Attribute Decision-Making

In the field of Information system and e-commerce, personalized recommendation is a very mature technology and performs well. However, the rapid growth of the amount of data not only makes it difficult to store and manage data, but also blocks the development of personalized recommendation as well as diversity. However, as a new computing model and service delivery model, cloud computing makes it possible to solve massive data management and improve the speed of recommendation. Besides, cloud computing also brings new problems to personalized recommendations.According to users’personalized demands, it’s the basic duty for personalized recommendation to recommend services to users, which could finish users’ demands. The existing methods for personalized recommendation of cloud service solve the problem of searching users’ service to some degree. However, these existing methods still have some defects. Firstly, although theses existing methods emphasizes the importance of personalized demands, they don’t describe user’s personalized demands apparently. Secondly, these existing methods don’t take the difference between the current demands and the non-current demands. Thirdly, users’ demands won’t stay in the same position and the demands will change with times on. However, most of existing methods don’t considering how to catch users’current demands. Thirdly, user won’t just focus their demands on the function of services. Comparing with the applications on PC, the real-time performance of applications on cloud platform will catch users’eyes, so user’s personalized demands includes demands on functions and performance. Finally, users usually have multiple demands on different aspects, but few methods take that into consideration. There seldom exists the methods which takes the above four problem into consideration. Therefore, this paper proposes a personalized recommendation model for cloud services based on Markov model and multi-attribute decision-making, hoping to solve the proposed problems to some degree. By analysis of users’habits of using the cloud service, the process of using cloud service is modeled as Markov model. In addition, aiming at the property of services, we propose a method to merge multi-way utility. By end of the paper, the results of experiment shows that the proposed method achieved the goal of solving the issues raised and improve the performance of personalized recommendation to some extent.