On Behavioral Dynamics Of Social Tagging And Its Application

The Social Tagging System (STS), such as Delicious, Flickr, CiteUlike, has been greatly promoting the development of the Internet in recent years and it has been getting increasing attention of the researchers. Based in the concept Web2.0, STS can make the information sharing more convenient. A noticeable issue of STS is the study of its intrinsic tagging mechanism, about which the previous study mainly stay on the macroscopic level of the distribution of the tags, focusing on their applications, while the reason or the mechanism which leads to numerous features of STS and its evolution law in the theoretical level still needs further investigation.Based on the analysis of the users’tagging behaviors, we propose a user-tagging-behavior model in accordance with the tagging behaviors of the STS users. With the idea of Agent-Based Modeling (ABM), we implement and validate the simulation model, through which we carry out deeper research into the intrinsic tagging mechanism of the STS. Furthermore, we explore the reason for the drawbacks of the existing STS and provide some suggestions to overcome them. Then, on the basis of the prior studies, we develop the social tagging system of an "Online Science and Technology Cooperation Community"(OSTCC), by incorporating the explored tagging mechanisms. For the one thing, we can make a primary validation of the tagging mechanism through the use of this system. For the other thing, the system can contribute to the construction of further tagging system as a reference application.The major contributions of this study can manly be summarized as follows.First, we build a user-tagging-behavior model considering the influence of both the individual user’need to organize their information resource and the impact on their tagging behaviors by the existing community-members’tags on their tagging behaviors with the improved Information Value Model and Preferential Attachment Model as the basis.Second, we implement the above model on the platform of JUNG (Java Universal Network/Graph Framework) under the concept of ABM to simulate the model and compare the simulation results with the actual delicious tagging data. Through such comparison, we verify the model and provide a reasonable intrinsic tagging mechanism. In the mean time, through the analysis of the evolution law of STS, we also discuss the defects of the existing STS and provided several suggestions for improvement.Third, we discuss how to apply the suggestions mentioned above to the construction of a practical and efficient STS for the OSTCC and designed a framework for building a STS. Furthermore, we develop a social tagging system prototype based on the platform of SpaceBuilder, a free software framework for social networking sites.In conclusion, from the theory aspect, we in this thesis propose a intrinsic tagging mechanism of the STS which can well explain the distribution of the existing social tags and provide a novel research idea for the study of the intrinsic mechanism and other related research issues; from the practical aspect, we discuss and analyze the reasons for the defects of the existing STS, and made some corresponding practical guidelines which are of certain practical significance for the construction of actual social tagging systems.