On team formation with expertise query in collaborative social networks
On team formation with expertise query in collaborative social networks
复制标题
DOI：
10.1007/s10115-013-0695-x
复制
发表时间：
2015-02
影响因子：
2.7
通讯作者：
Cheng-te Li;M. Shan;Shou-de Lin
中科院分区：
文献类型：
--
作者：
Cheng-te Li;M. Shan;Shou-de Lin
给定一个协作的社会网络和一个由一组所需技能组成的任务，团队组建问题的目标是找到一个专家团队，他们不仅满足给定任务的要求，而且能够以有效的方式相互沟通。本文扩展了原来的团队形成的问题，以一个广义的版本，其中的专家选择为每一个所需的技能也指定的数量。构建的团队需要包含足够数量的专家，以满足每项所需的技能。我们开发了两种方法来组成团队的proposedgeneralized团队组建任务。首先，我们考虑的具体数量的专家设计的广义增强Steiner算法。其次，我们提出了一种基于分组的方法浓缩的专业知识信息，一个紧凑的表示，组图，根据所需的技能。群图不仅可以减少搜索空间，而且可以消除冗余的通信成本，并在编译团队成员时过滤掉不相关的个体。为了进一步提高组合团队的有效性，我们提出了基于敏捷度的度量方法，并将其嵌入到所开发的方法中。在DBLP网络上的实验结果表明，该方法所组成的团队在有效性和效率方面都有较好的表现。
Given a collaborative social network and a task consisting of a set of required skills, the team formation problem aims at finding a team of experts who not only satisfies the requirements of the given task but also is able to communicate with one another in an effective manner. This paper extends the original team formation problem to a generalized version, in which the number of experts selected for each required skill is also specified. The constructed teams need to contain adequate number of experts for each required skill. We develop two approaches to compose teams for the proposedgeneralized team formation tasks. First, we consider the specific number of experts to devise the generalized Enhanced-Steiner algorithm. Second, we present a grouping-based method condensing the expertise information to a compact representation,group graph, based on the required skills. Group graph can not only reduce the search space but also eliminate redundant communication cost and filter out irrelevant individuals when compiling team members. To further improve the effectiveness of the composed teams, we propose adensity-basedmeasure and embed it into the developed methods. Experimental results on the DBLP network show that the teams composed by the proposed methods have better performance in both effectiveness and efficiency.