数据驱动环境下数据画像若干问题探析

Analysis on Several Problems of Data Profile in Data-driven Environment

[目的/意义]数据画像构建当前已引起产业界和学术界的高度关注。对大数据环境下数据画像基础理论的若干问题进行提炼与总结,可进一步丰富相关理论体系并指导数据画像应用。[方法/过程]在对现有数据画像实践与研究进行总结归纳基础上,对数据画像发展、概念、本质特征与分类、实现路径与通用方法等关键问题进行剖析,并对数据画像法律问题及标准体系建设等研究方向提出展望。[结果/结论]数据画像是对客体对象进行特征提取与分析的系统性工程,其本质是获得客体情报,其来自计算机科学、图馆书学情报学等领域对用户数据画像、城市数据画像等主题的探索。当前学界对用户画像等概念及其来源的认识仍然较为混乱,对数据画像尚未形成统一的认识。当前数据画像的主要研究领域聚焦于用户画像,逐渐向城市画像、实体画像等客体发展。数据画像的基本路径包含4个部分,且呈现方式不固定,包含表格、词云等多种呈现方式。[局限]对数据画像众多技术方法实现细节的分析总结不够。

[Purpose/significance] The construction of data profile has attracted great attention from industry and academia.Refining and summarizing some problems of the basic theory of data profile in big data environment can further enrich the relevant theoretical system and guide the application of data profile.[Method/process] On the basis of summarizing the current practice and research of data profile,we analyze the key issues of data profile development,concepts,essential characteristics and classification,implementation paths and general methods,and put forward the outlook of research on the legal issues of data profile and standard system construction of data profile.[Result/conclusion] Data profile is a systematic project for feature extraction and analysis of objective entities.Its essence is to obtain object information,which comes from the exploration of user profile and city profile in the fields of computer science,library and information and so on.The current academic circles are still confused about problems such as what is user profile and where does it come from,and the understanding of data profile has not formed a unified understanding.The current main research field of data profile focuses on user profile,and gradually develops into objects such as city profile and entity profile.The basic path of the data profile consists of four parts,and its presentation mode is not fixed,including multiple presentation modes such as tables and word clouds and so on.[Limitations] The analysis and summary are not sufficient enough about many technical method details of data profile.