Aiming at the problem that the influence of hierarchical relationships on interests failed to consider in the existing micro-blog recommendation, this paper proposed a micro-blog recommendation algorithm base on hierarchical social relationships. Firstly, it calculated the relative intimacy intensity and social growth rate based on the time window to construct a hierarchical network. Secondly, it quantified influence, attractiveness, and closeness of interaction in the network to tap the potential interest of target users. Then, it obtained the explicit interest of target users based on short text expansion strategy. Finally, it integrated potential interest with explicit interest, calculated the similarity between the fused interest and the micro-blog to be recommended to achieve recommendation. The experimental results show that compared with the classic micro-blog recommendation method, the new method is better than other methods in accuracy, recall, F-value, and MRR, with a maximum increase of 14.73%. The experimental results show that comprehensive consideration of potential interest and explicit interest can improve the effect of micro-blog recommendation.

[2021-12-05] Printed Article /doc/2021.06.0220/v2