认知计算视角下文献资源书目著录研究

Research on Bibliographic Description of Literature Resources from the Perspective of Cognitive Computing

认知计算是感知和思考信息处理过程的学科系统。认知计算系统从已有的书目数据中提取相关概念和个体关联,利用机器智能,创造独立于书目格式和以往学习经验之外的数据模型。新的书目著录模式,强调数据科学、认知体系与RDA编目规则的有机结合,分析认知计算系统中“数据层—信息层—知识层”几个层面与FRBR书目世界3组实体的关系,妥善处理编目工作中数据内容庞杂、信息动态分布与知识隐含层级等问题。借助数据挖掘、视觉认知与自然语言处理等认知手段,文献资源编目工作需要经历知识获取、知识编码、知识推理与知识验证4个环节,最终实现书目数据的信息化,为图书馆馆藏的合理布局、阅读推广的开拓创新与个性化信息服务的构建提供智能支撑。

Cognitive computing is a subject system that perceives and thinks about the information processing process.The cognitive computing system extracts relevant concepts and individual associations from existing bibliographic data,and uses machine intelligence to create bibliographic data models independent of bibliographic format or past learning experience.The new bibliographic description model emphasizes the organic combination of data science,cognitive system and RDA cataloging rules.By analyzing the relationship between“data layer-information layer-knowledge layer”in cognitive computing system and the three groups of entities in FRBR bibliographic world,the problems of complicated data content,dynamic information distribution and knowledge hidden level in document cataloging can be properly handled.With the help of cognitive means such as data mining,visual cognition and natural language processing,the cataloging work of literature resources requires four steps:knowledge acquisition,knowledge coding,knowledge reasoning and knowledge verification.Finally,the informationization of bibliographic data will eventually be realized,which will provide intelligent support for the rational layout of library collection,the innovation of reading promotion and the production of personalized information service.