【作者】 张素芳；

【导师】 王知津；

【作者基本信息】 南开大学 ， 图书馆学， 2012， 博士

【摘要】 科研人员学术创新的一个重要来源是学科的融合与交叉，面对海量数据，其交流趋向于从同事/团队、会议、论文等传统的正式交流渠道转向SNS、博客、论坛等网络非正式交流渠道。近年来，社交网络在科研中的应用增长呈上升趋势。学术科研类社交网络依托于已有的科研机构、出版商，或者从原来的研究性BBS、个人学术博客等发展而来，从最初的邮件发展到现在的学术空间，呈现迅猛的发展态势。据科学堂学术导航不完全统计，目前国内的学术科研类论坛相关网站已达上百个。如Peerage of Science系统、学问社区、CNKI社区等。其中哈佛的ResearchGate已经拥有130多万用户，而Academia.edu目前已经共享了110余万论文，每月访问量超过三百万人。科研社交网络发展势头良好。在社交科研网络中，科研人员的关联方式主要与其所属机构、学科、交流制度和学术活动相关，跨越地域、学科、身份、种族、机构等限制的弱关联科研人员之间的多向、动态、随机、随时、随地的沟通和交流对于科研创新更为有效。为了促进科学交流和科研创新，需要对网络社区中的各种学术资源之间的关联进行揭示和管理，发现关联规律、形成关联模型、明确关联网络的运行机制和关联的实现方案。本研究旨在帮助科研人员了解网络社区的功能和关联，为科研人员的学术研究需求的满足和找到同道者提供一个线索，也能为相关社交网络的功能设计提供一个理论参考，为企业和研究机构开发学术类网络社区产品和服务提供参照。研究基于社会网络理论、复杂网络理论、关联挖掘理论等开展对网络社区学术资源关联模型和机制的研究。论文以社交网络中的各类科研人员为研究对象，分析其学术信息需求和行为特点，为根据这些特点构建学术资源关联互动模型提供了基础。基于上述需求与行为分析，结合复杂网络理论和系统论的观点，构建了网络社区的学术资源关联模型，将网络社区的学术资源关联分成两大类型，即信息和人际网络。二者在信息和网络技术的支撑下围绕网络信息行为互动呈现螺旋上升趋势。随着时间的推移和信息技术的发展，网络社区中的信息关联和人际网络关联被信息行为推动着从原生状态发展到次生状态，再到复杂状态，其类型、范围、复杂程度、动态性和综合性都不断发展。为了验证所提模型的正确性，采用网络分析法对科学网博客社区图书馆学情报学领域的所有博客进行了实证分析，以发现该博客社区中纯好友网络和基于互动的好友网络的不同；然后又以新浪图林博客圈为基础，选择了大约60个博客样本，分析其基于tag共现的博主网络的结构和特征。实证结果验证了前面提出的理论模型。为促进网络社区的健康发展，对网络社区学术资源关联网络的运行机制进行了探讨，并提出了网络社区学术资源关联的实施方案。研究发现，网络社区学术资源关联网络的发展符合群体动力学观点，受到来自群体意见领袖、群体成员、群体规范等因素的影响，本文对其中意见领袖的形成于挖掘、社区成员的参与与激励、科研兴趣的分析与挖掘进行研究，认为三者共同作用于学术资源关联网络的发展。为保证网络社区的学术资源关联能得以顺利实现，分别从分类法、主题法和信息质量控制保证等角度对相关理论方法和技术进行了分析研究。本文的创新之处表现为以下几点：第一，构建了网络社区中的学术资源关联模型。技术的进步支撑和推动了社会网络中人际网络和信息网络的发展。人际关联和信息关联是学术资源关联的主要表现形式，通过网络社区的互动，这种关联会发生变化，形成更为复杂、紧密、多样化和多类型的关联体系；第二，提出了网络社区学术关联发展的动力机制。社区的学术意见领袖对于引领社区的话题导向和推动社区的交流、活跃度起着积极的作用；每个社区成员畅所欲言、有选择性参与社区的话题和活动，形成了推动社区人际网络和信息网络发展的系统自组织力量；技术的发展是社区得以发展的物质保证。第三，提出了网络社区关联信息组织的实现方案。基于人际网络的关联分析和基于知识发现的数据挖掘，从分类、主题和质量保证三个方面论述了网络社区中学术资源关联组织的实现保障，为网络社区中的信息组织的发展和完善提供了一个可操作化的应用框架，这也是对网络社区信息组织理论和实践研究所做的一种有益的探索和尝试。更多还原

【Abstract】 One important source of academic research innovation is the mixture and cross of different disciplines. Faced with enormous data, academic research communities shifted from physical and normal ways---colleague discussions, meetings, papers, et al.---to such virtual and informal ways as the Internet, BBS, Virtual Communities, and so on. In recent years, academic social network had developed rapidly, and social network had been applied in the academic fields for much wider usage. According to the statistical data from Datatang, a famous sharing platform in China, there have been more than100academic social networks in China, and the number is still in growing. Some famous academic social network sites are listed but not limited as the Peerage of Science, CNKI community, ResearchGate from Harvard University, Academia.edu, and so on. With the development of Web2.0, the organization methods of information had changed dramatically. Different from the traditional authority-controlled organization mode, it had become popular that each member of the network community commits his own wisdom and energy to the community in the network context. In the social network, academic researchers had correlations with others in many ways, such as affiliations, disciplines, communications system, and academic activities. Weak ties that cross over the physical areas, disciplines, ranks, status, nationalities limits have shown their priorities in the network communication, because such kinds of communication ways seemed to be more instructive and thus more effective for academic innovation.In order to effectively organize and reveal the correlation information that are widely existed in the network communities, and to make the organization methods of information more suitable for the academic innovation requirements, it’s necessary to probe into the organization mode of network academic communities. The result of the research is also useful for the product and services design for community-related enterprises.The research was conducted based on the theory of social network, the complex network theory, association mining, and so on. Different kinds of academic researchers are considered as a whole to analyze the characteristics and their academic information requirements and behaviors, organization mode of the relevant information are constructed, and the control mechanism of information organization was analyzed, which formed the theoretical and methodological base of the research.After analyzing the influence of the social communities on the academic communications, the characteristics of information requirement and the mode of information behavior were summarized, forming the foundation of the organization mode being constructed in the next chapter. There were two flows in a network community, one was the information flow, and the other one was human relationship flow. The theory framework in a whole was constructed based on this concept. The organization of community information was viewed as a two-layered net, with the information resources network laid under the human relationship network. The information flow circling in the information network and the relationship flow lying in the social communities interweaved with the information activities. The two networks took on an upward spiral pushed by the other force----network technology. Along with the development of time and information technology, the information linkage and human relationship were pushed from primary state to secondary state and sophisticated state finally. The three kinds of states were differed in types, range, complexity, dynamics and integrity.In order to test the theoretical mode presented above, empirical study was conducted, taking the academic information blogosphere of Sciencenet and Sina for examples. The experimental result verified the theory.In order to organize the community information in an effective, convenient and high-quality way, it was suggested that the role of academic community leader, member participation, relationship construction, the discovery of academic interest, and the control on the academic information quality be considered.As for the implementation of correlations, traditional information organization methods such as classification and subject analysis were remained useful for organizing information in the communities. Furthermore, these methods had been developed to be suitable for the new objects. Social collaborative tagging were described and analyzed as the new organization method in social network.The innovations of the thesis were as following:first, the research extended and enriched the dimensions of relationship analysis on the basis of the existing research. Secondly, the thesis built an integration model for academic resources relationships in network communities, and found the spiral rising trend of the academic resources. Thirdly, the thesis put forward the running mechanism and implementation strategies of the academic resources correlations. The conclusions of the thesis had some guiding function for the practice of enterprises which had been intended to construct or manage network communities.更多还原