CAJ | 학술논문

对高效的个性化PageRank算法进行综述，从个性化程度、可扩展性、计算效率和精确度等方面对部分个性化算法、可扩展的PPV算法和混合算法等3类算法进行了详细分析和比较，并讨论了目前工作中的不足及未来的研究方向。
대고효적개성화PageRank산법진행종술，종개성화정도、가확전성、계산효솔화정학도등방면대부분개성화산법、가확전적PPV산법화혼합산법등3류산법진행료상세분석화비교，병토론료목전공작중적불족급미래적연구방향。
As personalized PageRank has been widely leveraged for ranking on graph-structured scenarios, its computation efficiency becomes a prominent issue. We survey on an array of work that concentrate on efficient and scalable personalized PageRank computation, ranging from earlier work that attempt to use partial precomputation to improve online efficiency, to recent work that estimate approximate PPV for full peronalization and the hybrid methods. We compare these methods in terms of the ability of personalization, scalability, online/oitline efficiency and accuracy. We also point out a few possible research directions at the end of this paper.