【作者】 李洁；

【导师】 毕强；

【作者基本信息】 吉林大学 ， 图书情报与档案管理， 2019， 博士

【摘要】 我们已经从信息时代走进了数据驱动的“智能时代”,数据成为人们认识和解决问题的新的逻辑起点。“数据驱动”打破了基于知识解决问题的思维桎梏,形成了从问题到数据又回归问题的新方法论认识——基于数据解决问题。这一研究范式将数字图书馆知识发现服务从对问题本源的探索推向知识服务的本真,可以说,从数据直面用户、管理和服务为大数据环境下的数字图书馆知识发现服务供给侧的改革提供了一种新思路:知识发现服务要改变的不只是管理技术、管理规则或服务形式,而要涉及整个管理理念和服务体系。而大数据环境中,数字图书馆信息发生源越来越多,数据产出量越来越大,数字资源增长速率越来越快,数据异构性越来越明显,数据老化节奏越来越快,低价值密度隐患的知识饥渴和数据海啸的矛盾越来越突出,用户对发现服务的需求越来越多元,数字图书馆数据资源正面临着重新被发现的挑战。迎面变化和挑战,数字图书馆的知识发现服务不单要完成从文献数字化到内容数据化的知识组织转型,更应实现数字资源从内容数据化到数据智能化的价值开发和智慧洞见。数据驱动的科研范式开辟了知识发现的新路径,开启了数字图书馆知识服务的时代新转型。探索数据驱动理念下的数字图书馆知识发现服务模式的新形态,需要学习、内化数据科学相关理论,需要剖析知识发现的驱动要素和作用机制,需要打破传统的资源发现固化模式,创建知识发现服务的创新生态功能圈。融合数据驱动和知识发现的双重技术优势,数字图书馆知识发现服务创新模式应趋从数据化、数据向知识转化的语义关联、可视化和智能化驱动维度寻求用户数据、内容资源数据、专家数据、业务数据的新协同,开发用户画像、研究设计指纹、精准文献推荐等的新应用,强化数据的集群整合、提升平台的绿色联通、实现用户界面的友好交互,使数字图书馆成为支持用户知识探索与发现创造的智能服务系统,使数据资源最大化的进行价值开发与知识转化,使用户随时随地都能受益于数字图书馆高效、便捷、友好与智能的知识发现服务体验。基于此,本文通过对数据驱动、知识发现研究成果的追本溯源,界定数据驱动下的数字图书馆知识发现服务的核心理念;通过文献分析、调查访谈、仿真实验、模型训练等方法的综合运用,分析数字图书馆知识发现服务创新的数据环境、驱动机制、创新模式、模式应用以及创新策略制定。围绕主要研究内容,本文第三章从数据环境特征、数据环境变化和数据环境开发分析数据驱动下的数字图书馆知识发现服务的机遇与挑战;第四章结合数据要素、数据驱动过程、数据驱动维度探讨数字图书馆知识发现服务的数据驱动动力机制、流机制、协同驱动机制和数据驱动控制机制;第五章通过对数字图书馆知识发现服务模式创新衍变的内在使命分析,指出数字图书馆知识发现服务创新模式的构建依据、构建基础和构建过程;第六章对数字图书馆知识发现服务创新模式进行具体的用户画像、研究设计指纹、文本推荐和多粒度检索决策应用;第七章针对数字图书馆知识发现服务创新模式的具体瓶颈给出各驱动维度的应对策略。具体内容阐述如下:第3章数据驱动下数字图书馆知识发现服务的数据环境分析本章是对大数据驱动环境下的数字图书馆知识发现服务场域的情境解构。首先,基于大数据的4V特征,面向全数据,分析数字图书馆知识发现服务在数据形态、存在方式、存储模式、存储内容、数据价值等方面的特性。其次,探讨数据化、新一代信息技术、数据分析思维、数据密集型科学发现范式影响下的数字图书馆知识发现服务革新的优劣利弊。最后,基于环境特性和环境变化的双向作用状态定位数字图书馆知识发现服务发展的开发方向。明确本文研究目的的同时,引出4、5、6、7章节的主要研究任务。第4章数字图书馆知识发现服务创新的数据驱动机制分析本章作为第5章的铺垫,详细解析数字图书馆知识发现服务平台的数据要素和驱动作用形式。通过用户数据、资源内容数据、专家数据的分类界定,为第6章科研用户画像、研究设计指纹、精准文献推荐等的服务模式应用提供数据基础;通过数据化、语义关联、可视化、智能化的数据驱动维面的层级解构,为第7章的创新策略制定奠定优化主线;基于数据要素、驱动过程和驱动维面,从内外力作用的动力机制、输入-输出的流机制、数据融合的协同驱动机制以及数据驱动控制机制具体呈现数据驱动与知识发现服务交互融合的催化反应。第5章数据驱动下数字图书馆知识发现服务创新模式研究在前文研究的基础上,本章首先对数据驱动下数字图书馆知识发现服务创新模式的构建进行内在逻辑分析;其次,从资源发现既有模式、知识产品和技术支持方面阐述实现数字图书馆知识发现服务创新的外在基础;最后,综合内在逻辑和外在基础,进行创新模式的基础框架和平台架构的初步解构,并在此基础上进行数据驱动下的数字图书馆知识发现服务创新功能圈构建。第6章数据驱动下数字图书馆知识发现服务创新模式应用研究本章在第5章提出的创新模式的基础上,分别利用科研用户数据进行数字图书馆百度发现的科研用户画像构建,利用文献数据进行以研究对象、研究问题与研究方法为核心要素的研究设计指纹构造,结合用户画像和研究设计指纹实现精准文献推荐,并通过用户检索实验验证多粒度检索决策的优势。第7章数据驱动下数字图书馆知识发现服务创新策略研究基于第4章对数据驱动维度和驱动机制的分析,本章旨在明确数据化、语义化、关联化、可视化和智能化驱动的创新方向并进行相应的优化路径设计,针对数字图书馆知识发现服务创新的制约因素,给出切实可行的解决方案与对策建议。大数据环境下,重新界定数据驱动下的数字图书馆知识发现服务的内涵、探讨数字图书馆知识发现服务的数据驱动机制、创新数字图书馆既有的资源发现服务模式,有利于从方法论认识层面为数字图书馆知识发现服务的供给侧改革提供理论支持。数字图书馆知识发现服务的意义不仅在于它的统一检索及其延伸功能,更在于此基础上辅助科学发现的循证决策、智能管理和知识再造的服务价值。在人类不断探索未知与努力认识未知的道路上,数据驱动+知识发现的催化反应为科学发现的方法探索提供了一个可行参考,推动着数字图书馆知识发现服务在不断革新的历程中惠及更多的求知受众。更多还原

【Abstract】 We have entered the data driven intelligence age from the information age,and data has become a new logical starting point for people to understand and solve problems."Data-driven" breaks the shackles of solving problems based on knowledge and forms a new understanding of methodology that use data to solve problems.This research paradigm pushes the digital library knowledge discovery services return from the exploration of the root of the problems to the true nature of the knowledge services.Directing to the users,managements and services from data,so to speak,provides a new idea for the reform of the supply-side of knowledge discovery services in digital libraries in big data environments.Knowledge discovery services should not only change the management technologies,management rules or service forms,but also involve the entire management philosophies and service systems.But in the big data environments,there are more and more sources of digital library information,the data output is getting larger and larger,the growth rate of digital resources is getting faster and faster,the data heterogeneity is more and more obvious,and the data aging is more serious.The contradictions between the hidden dangers of knowledge hunger because of low-value density and data tsunami are becoming more and more prominent,and the demands of users for discovery services are becoming more and more diversified.The data resources in digital library are facing the challenge of being rediscovered.About the oncoming changes and challenges,the knowledge discovery services of digital libraries not only need to complete the transformation of knowledge organizations from document digitalization to content datamation,but also realize the value exploitation and wisdom insight of digital resources from content datamation to data intelligence.The research paradigm of data-driven opens up a new path for knowledge discovery and opens up a new era of knowledge services of digital libraries.Exploring the new form of the digital library knowledge discovery service models based on the data-driven requires learning and internalizing the relevant theories of data science.It is necessary to analyze the data-driven factors and the mechanism of knowledge discovery.It is necessary to break the rigid models of traditional resource discovery and create an ecological function circle of service innovation.Combining the dual technological advantages of data-driven and knowledge discovery,the innovation models of knowledge discovery services in digital libraries should seek the new synergy of user data,content resource data and expert data from aspects of datamation,semantization,association,visualization and intelligentization.And exploit new applications such as the user portraits,research design fingerprints,and accurate document recommendations,strengthen data cluster integration,enhance the green connectivity of platforms,achieve user-friendly interaction.Make digital libraries an intelligent service system that supports the explorations,the discoveries and creations of knowledge for users,maximize data resources for value developments and knowledge transformations,and enable users to benefit from digital libraries’ the efficient,convenient,friendly and intelligent knowledge discovery service experiences.Therefore,this paper defines the core concept of data-driven digital library knowledge discovery services through tracing the research results of the data-driven and the knowledge discovery.With the comprehensive application of literature analyses,surveys and interviews,simulation experiments,model trainings and other methods,analyze the data environments,driving mechanisms,innovation models,model applications and innovation strategies of digital library knowledge discovery service innovations.Focusing on the main research contents,the third chapter of this paper analyzes the opportunities and challenges of knowledge discovery services driven by data from the characteristics of data environments,data environment changes and data environment developments.The fourth chapter combines data elements,data-driven processes,and data-driven dimensions to explore the data-driven dynamic mechanism,the stream mechanism,the synergetic driven mechanism and the data-driven control mechanism of digital library knowledge discovery services.Chapter Five points out the construction requirements,construction foundations and construction processes of the digital library knowledge discovery service innovation models by analyzing the inherent missions of the innovations and evolutions of the knowledge discovery service models of digital libraries.Chapter Six realizes specific applications,such as user portraits,research design fingerprints,text recommendations,and multi-granularity retrieval decision-makings for the digital library knowledge discovery service innovation models.Aiming at the specific bottlenecks of the digital library knowledge discovery service innovation models,chapter seven provides the solution strategies for each driven dimensions.The specific contents are explained as follows:Chapter 3 Data-Driven Data Environment Analyses of Knowledge Discovery Service in Digital LibraryThis chapter is a contextual deconstruction of the digital library knowledge discovery service domain in big data driven environments.Firstly,based on the 4V characteristics of big data,this chapter from the perspectives of the all-data oriented analyzes the characteristics of digital library knowledge discovery services in terms of data states,existence modes,storage modes,storage contents and data value.Secondly,this chapter discusses the advantages and disadvantages of digital library knowledge discovery service innovations under the influence of datamation,new generation information technologies,data analysis thinkings and data-intensive scientific discovery paradigms.Finally,based on the two-way state of environmental characteristics and environmental changes,the development directions of digital library knowledge discovery services are positioned.While clarifying the research purpose of this article,the main research tasks of chapters 4,5,6,and 7 are drawn.Chapter 4 Analyses of Data Driven Mechanisms of Knowledge Discovery Service Innovation in Digital LibraryAs paving the way for chapter 5,this chapter analyzes the data elements and driving forms of the digital library knowledge discovery service platforms in detail.Through the classifications of user data,resource content data,and expert data,provides data foundations for service model applications such as research user portraits,research design fingerprints,and accurate document recommendations of Chapter 6.Through the hierarchical analyses of datamation and semantization,association,visualization and intelligentization,lays the optimization main lines for the formulations of innovation strategies of Chapter 7.Based on the data elements,the driven processes and the driven dimensions,present the interactive catalytic reactions of data-driven and knowledge discovery services from the analyses of dynamic mechanisms of the internal and external forces,the input-output flow mechanism,and the data-fusion synergetic driven mechanism and the data-driven control mechanism specificallyChapter 5 The Research on Data Driven the Innovation Models of Knowledge Discovery Service in Digital LibraryOn the basis of the previous researches,this chapter firstly analyzes the internal logics of the data-driven innovation models of digital library knowledge discovery services.Secondly,this chapter from the external bases,expounds the realization of digital library knowledge discovery service innovations from the aspects of resource discovery existing models,knowledge products and technical supports.Finally,the internal logics and external foundations are integrated,the initial deconstructions both of the innovation model’s basic frameworks and the architectures of platforms are carried out.And on this basis,the data-driven digital library knowledge discovery service innovation function circle is constructed.Chapter 6 Application Researches of Data-driven Knowledge Discovery Service Innovation Models in Digital LibraryBased on the innovation models proposed in Chapter 5,this chapter uses the research user data to construct the research user portraits of Baidu discovery in the digital library,and uses the literature data to construct the research design fingerprints with research objects,research problems and research methods as the core elements.And combined with user portraits and research design fingerprints to achieve accurate literature recommendations,and verify the advantages of the multi-granular retrieval decisions through user retrieval experiments.Chapter 7 Researches on Data-driven Innovation Strategies of Knowledge Discovery Service in Digital LibraryBased on the analyses of data-driven dimensions and driving mechanisms in Chapter 4,this chapter aims to clarify the directions of innovations in datamation and semantization,association,visualization and intelligentization,and to design corresponding optimization paths for digital library knowledge discovery services.Based on the innovative constraints,this chapter gives practical solutions and countermeasures.Under the big data environments,redefining the content of digital library knowledge discovery services driven by data,exploring the data-driven mechanisms of digital library knowledge discovery services,and the existing resource discovery service models of innovative digital libraries are beneficial to provide theoretical support for the supply-side reforms of knowledge discovery services in digital libraries from the perspective of methodologies.The significances of the digital library knowledge discovery services lies not only in their unified retrievals and their extension functions,but also on the bases of the evidence-based decision-makings,intelligent managements and knowledge re-creation services value of scientific discoveries.On the road of human beings constantly exploring the unknown and trying to understand the unknown,the catalytic responses of data-driven +knowledge discovery not only provide a feasible reference for the explorations of scientific discovery methods,and also promotes the knowledge discovery services of digital libraries in the process of continuous innovations to benefit more knowledge-seeking users.更多还原