面向科学社会计算的数据组织与建模方法

Data Organization and Modeling for the Social Computing of Science

科学系统的动态性与开放性使得针对系统内部海量多源数据的组织与建模难度不断增强,进而提升了科学系统研究的技术壁垒。为延伸科学社会学对科学系统的认知,本文强调科学知识对科学系统运行的影响,将科学视作一个以科学共同体为主体、知识流动为动力的社会系统,探索性地提出科学社会计算的概念,并探究面向科学社会计算的数据组织与建模方法。首先,通过构建三元链路模型,以厘清科学社会内部实体与数据构成,实现对科学社会的抽象化表示;然后,采用多源数据对科学社会中的实体进行数据化表征,并根据三元链路模型进行"实体—属性—关系—数据"的关联,实现科学社会数据建模;最后,以科学社会人际网络和知识扩散研究为例,阐释基于多源数据开展科学社会计算的过程与研究优势。本研究旨在在底层数据与计算技术之间搭建桥梁,为利用多源数据全景化剖析科学系统构成以及开展科学社会计算研究提供方向性的指导。图4。表1。参考文献40。

Science is a complex dynamic and open system which enhances the technical barriers to organize the massive a nd multi-source data for revealing the operation mechanism of the science system.Extending the definitionof the science system in the sociology of science this study emphasizes the role of scientific kno wledge.The science system is regarded as a scientific societywhere the scientific community works as themain participants and the diffusion of existing knowledge is the driving force to facilitate scientific creationan d production.Based on the idea the concept of"social computing of science"is proposed to generalizethe research on the theories and methods that focus on the complex changeable phenomena and issue s in the scientific society.Toassist the social computing of science the general framework of data organization and modeling is generated.First we propose a ternary link model to profile the operating process of the scientific society.The model consists of four types of entities i.e.the participants the locations the activities and the knowledg ethat connected the other three components.Through the connection among the entities the operationmode and potential laws of the scientific society can be more comprehensively reflected and can ass ist in modeling the scientific society from multiple perspectives.Then based on the ternary link model weanalyze the path of data o rganization by summarizing the related data and attributes of th e four entities.The extraction of the entitie’s attributes is regarded as a classification task based on whichthe data fusion method of theentities is developed.Finally from the perspective of the multi-layer network we explore the relations hip among different types of entities and propose the relationship integration model for the scientific society.Taking the cases on the interpersonal networks analysis and the cross-community knowledge diffusi onas examples the paper illustrates the process and advantages of social computing of science that is based on the fusion of multi-source data.Theoriginality of this study mainly consists of 1 A ternary link model is proposed to conceptualize the comp onents and their relationships in the scientific society which emphasizes the influence of knowledge on otherel ements.2 The two research scenarios visualize the concept of s ocial computing of science and provideideas for the specific appli cation of the proposed data organizationand modeling method.3 This study can build a bridge betweenthe underlying data an d computing technology and provide directional guida ncefor using multi-source data to comprehensivelyanalyze the composition of scientific society.4 figs.1 tab.40 refs.