Master by Research
School of Computer Science and Software Engineering, Faculty of Informatics
Yuan, Shaojie, Discovery of knowledge collaborative communities for multi-domain problem solving, Master by Research thesis, School of Computer Science and Software Engineering, Faculty of Informatics, University of Wollongong, 2010. http://ro.uow.edu.au/theses/3351
The rapid technological advancement in the modern world has brought about a surge in the quantity of available data. There are vast potential knowledge and predictive pattern that exists within the abundance of data. However the technique to locate desired information and to identify useful patterns still poses a problem in today’s information age . With the endowment of knowledge, many multi-domain problems arise, and can only be solved with diverse expertise. Hence the ability to obtain and utilise the knowledge within the myriad data to help solve multi-domain problems remains an important and challenging research issue. This thesis aims to find an approach to discover a knowledge collaborative community to solve a multi-domain problem by study transaction data. In this thesis, firstly, a core-based node ranking approach is proposed, which could be used for an expert finding task; secondly, a Knowledge Collaborative Community (KCC) approach to discover a group of experts to efficiently solve a multi-domain problem in a small-size or medium-size network is proposed; thirdly, a two-step KCC approach used in large-scale networks to discover KCCs is presented. The major contributions of this thesis are as follows. 1. A core-based node ranking approach used in event-based social network is proposed. The approach can rank nodes based on the importance and the activeness of the nodes in a network. The ranking result could vary based on different unit core (which demonstrates different demands). This approach could be deployed to rank experts. 2. A Knowledge Collaborative Community (KCC) approach is proposed, which could be used to discover KCCs in small or medium networks. Although some existing work proposed similar approaches, most of them considered only part of factors which might influence knowledge collaboration. In KCC approach, more factors, such as knowledge coverage, size of community, personal desires, are considered. Furthermore, the knowledge level of an expert is introduced as an important factor which may impact the efficiency of knowledge collaboration. 3. A two-step KCC approach is proposed in this thesis to discover KCCs in largescale networks. Compared with previous research, this approach has better performance in poorly connected and/or decentralized large-scale networks.