Abstract:

[Background] China Science and Technology Cloud (CSTCloud) has been a widely concerned cloud platform which aggregates resource of cloud computing, cloud storage and high performance computing for scientists since its establishment in April, 2018. As the entrance of aggregated resources, CSTCloud is of great importance for users to obtain and utilize resources. [Methods] In order to understand user activities and stickiness, this paper develops an improved RFM model to evaluate loyalty of users based on real data of CSTCloud visitors. [Results] The model reflects the user data characteristics and the negative correlation between Recency index and the user value. In addition, users are classified by real loyalty analysis results, assessing the potential value of users for CSTCloud. [Conclusions] Based on real data, this study can help the CSTCloud operator to learn about user activity characteristics and to provide personalized services for classified users.

Key words: CSTCloud, user activity, user loyalty, RFM model