Research On Team Formation In Social Network

Nowadays, with the development of social diversity, people need to acquire more and more social information and share more resource. Therefore, it makes us to build a collaborative team to complete a given task together. A given project which is completed successfully requires a set of skills and our goal is to find a set of experts that own minimal communication cost among them to complete it effictively.We call that team formation problem. Team formation problem has become a hot research in social networks. But there is a common problem that using 0-1 form to measure whether an individual has a skill in the existing research. Obviously, it is unable to define individuals whether have skills in practice,for example, individuals may be quite familiar with certain skills or just know a little. The way to define whether an individual has certain skill will largely influence the efficiency of the team to complete the task .Moreover, in the present study, it only focuses on considering the number of past finishing tasks while not deeply digging social relationships between individual categories as well as the individual's own other properties in calculating the strength of relationships between individuals in social network. These largely influence social relationship and the team cooperation between individuals.To solve the above problems, we put forward the methods to generate the team and finally form a highly efficient team. The main research work of this paper includes the following content:Firstly, we give problem definition and related concepts as well as contribution measure of individual's skills and a model of relationship to calculate the relationship between individuals.Secondly, according to the different forms of team, team formation problems can be divided into: without a leader of team formation problems and with a leader of team formation problems; In without a leader of team formation problem,we use the sum of minimum steiner tree path as the team communication cost. we propose a Minimum Covering Steiner-based Team Formating Algorithm to find the best team and a method which is Minimum Set Covering-based Team Forming Algorithm to search for the best team. In the study of with a leader of team formation problem,we use the minimum distance as a team leader of communication cost, and proposed Minimum Leader Distance based Team Formating Algorithm to reduce the search space and improve the efficiency of searching; finally we found the best leader and the best team.Finally, we use Github dataset and the DBLP dataset to experiment to verify the feasibility and effectiveness of all of the above methods as the same as compare with existing algorithms from the following fields: team communication cost, members of the team, team connectivity and scalability of algorithms .we find that our algorithms has better feasibility and effectiveness.