基于个人知识地图的专家推荐

摘要

利用维基百科作为背景知识构建出专家个人的知识地图,从而直观量化地度量专家的知识构成和研究兴趣.在此基础上,提出了基于知识节点密度和最大公共子图的2种推荐算法,并且将这2种推荐算法和经典的推荐算法结合.最后,用一个真实的数据集合验证:这种基于个人知识地图的专家推荐是有效的.%The personal knowledge map is constructed for every expert by leveraging the knowledge from Wikipedia, and thus the experts' knowledge can be measured precisely and directly. Based on the personal knowledge map, two recommendation algorithms are proposed, and the two algorithms can be combined with classic algorithms lineally. Finally, the experiments on a real dataset demonstrate this approach is effective.