![](/Areas/202005130001/Content/ZHYJY/img/niu1.png)
![](/Areas/202005130001/Content/ZHYJY/img/niu2.png)
![](/Areas/202005130001/Content/ZHYJY/img/niu3.png)
![](/Areas/202005130001/Content/ZHYJY/img/niu4.png)
英文标题: | A New Paradigm of Combining Big Data and Survey Data Based on the Theoretical Perspective |
摘要: | 本文以人脉圈层研究为例,将抽样调查得到的扎根真相与在中国广泛使用的一款社交软件A的大数据结合,建立人脉圈分类模型。在理论、数据挖掘、回归模型和分类预测模型及其解释工具的对话中,通过一次次抽样取得扎根真相,进行一轮轮的模型校准,发展出越来越精准的预测模型。本案例展示的大数据与结构化数据整合的研究范式是社会科学理论导引下的大数据研究方法论的实践。 |
英文摘要: | When big data based on social network, social media, and sensors is combined with ground truths collected through qualitative or quantitative social survey methods in social sciences, a new research method is created based on the combination of big data and structural data driven by social theories. Using traditional survey methods to collect small-scale ground truths makes it possible to combine social science research questions, ground truths with existing big data, to establish a prediction model that provides a triangular reciprocal dialogue for data mining, theoretical modeling with causality inference, and prediction algorithm. Taking the research of the guanxi circle model as an example, this paper combines the ground truths obtained from the survey data with the big data of a widely used social network software in China to build a guanxi classification model. In the dialogue between theory, data mining and prediction algorithms, increasingly accurate prediction models are developed and a measurement of tie strength is created after multiple runs of sampling and model modification. |
作者: | 罗家德、高馨、周涛等 |
作者单位: | 清华大学社会学系与公共管理学院(罗家德、高馨);电子科技大学计算机科学与工程学院、大数据研究中心(周涛) |
期刊: | 社会学研究 |
年.期:页码 | 2021.2: |
中图分类号: | |
文章编号: | |
关键词: | 大数据;扎根真相;人脉圈分类模型;社会计算学 |
英文关键词: | |
项目基金: | 感谢清华大学社会网络研究中心以及清华大学校内自主科研项目“基于通讯数据的关系强度与社会资本挖掘”(计划编号:2016THZWYY03)、腾讯研究院“基于QQ大数据意见领袖识别研究”(项目编号:20182001706)的资金支持。 |
欢迎阅读《社会学研究》!您是该文第4072位读者! |