虚拟与现实:电子踪迹大数据质量与知识发现

Virtual and Reality: Inquiry on the Quality of Digital Trace Data and the Knowledge Discovery

电子踪迹数据一般是结构化/可结构化、量化/可量化的,相对易于分析,数据处理精度一般较高,因此受到数据驱动型知识发现的重视。然而,受商业利益、政治权力等因素的影响,很多电子踪迹大数据的真实性、客观性、自然性、准确性存疑。网络信息空间具有记录空间和行动空间双重属性,加之其越来越强的工具性特征,使它难以真实完整地映射现实社会空间的状况。社会科学研究者在使用电子踪迹大数据时,须审慎考量数据质量。

The digital trace data is generally structured/can be structured, quantitative/quantifiable, so its analysis is relatively easy, and the data processing accuracy is generally higher. Therefore, the digital trace data at the present stage is especially valued by data driven knowledge discovery. However, due to factors like commercial profit and political power, the quality of many sorts of digital trace data is doubtful. The network information space in essence is recording space and also is action space, especially its more and more obvious instrumental characteristics, which make it impossible to reflect the social reality space factually, accurately. While applying these data to scientific research, social scientists should carefully consider the issue of data quality.