在社交网络中执行去中心化搜索的算法
我想找出所有现有的利用社交网络结构特性的去中心化算法。到目前为止，我知道以下算法 -
1）最佳连接搜索 - Adamic 等人
2）随机游走（不利用任何结构属性，但仍然是分散的）
3）汉明距离搜索
4）弱/强关系搜索
5）余弦相似度搜索 (CCS)
6) 信息气味搜索 (ISS)
如有任何帮助，我们将不胜感激
I want to find out all the existing decentralized algorithms that exploit the structural properties of social networks. So far I know the following algorithms -
1) Best connected search - Adamic et al
2) Random Walk (does not exploit any structural property but still it is decentralized)
3) Hamming distance search
4) Weak/Strong tie search
5) Cosine Similarity Search (CCS)
6) Information Scent Search (ISS)
Any help would be appreciated
如果你对这篇内容有疑问，欢迎到本站社区发帖提问 参与讨论，获取更多帮助，或者扫码二维码加入 Web 技术交流群。
绑定邮箱获取回复消息
由于您还没有绑定你的真实邮箱，如果其他用户或者作者回复了您的评论，将不能在第一时间通知您！
发布评论
评论（2）
摘自论文“搜索社交网络专业知识：模拟
潜在策略”：
广度优先搜索（BFS）向所有邻居广播查询，而不是根据启发式选择邻居。它可以找到最接近源的目标，但带宽成本极高（如 p2p 文件共享）余弦
相似性搜索 (CCS) 通过将汉明距离除以邻居拥有的出度关系（朋友）总数来减少高度影响。
信息气味搜索 (ISS) 选择下一个具有最高匹配度的人。我们的算法实现与 Yu 和 Singh 略有不同，因为我们需要使他们的算法适应安然数据集，而我们使用自动生成的关键字配置文件。
From the paper "Searching For Expertise in Social Networks: A Simulation
of Potential Strategies":
Breadth First Search (BFS) broadcasts a query to all of one’s neighbors instead of picking a neighbor according to a heuristic. It can find the target closest to the source but with extremely high bandwidth costs (as in p2p file sharing networks).
Cosine Similarity Search (CCS) decreases the high degree impact by dividing the Hamming distance by the total number of out-degree relations (friends) a neighbor has.
Information Scent Search (ISS) picks the next person who has the highest match score (which we call information scent) between the query and his profile. Our implementation of the algorithm is slightly different from Yu and Singh, since we needed to adapt their algorithm to the Enron data set. We use the automatic generated keywords profile instead.
[PDF] 复杂网络和去中心化搜索算法
http://www.cs .cornell.edu/home/kleinber/icm06-swn.pdf
[PDF] Complex networks and decentralized search algorithms
http://www.cs.cornell.edu/home/kleinber/icm06-swn.pdf