发现因果的算法：被忽视的计算社会学工具箱
Causal Discovery Algorithms: A Neglected Toolbox for Computational Sociology
-
摘要: 社会学者长期以来聚焦于因果识别的方法及其应用，却对因果发现的相关进展缺乏关注。在回顾因果推断两种理论传统的基础上，从算法原理、应用路径和方法关联三个方面，系统介绍因果发现的经典算法，探讨它们在研究实践中的应用方式，以及与其他计算社会学方法结合的可能性。面对大语言模型的崛起，因果发现算法的应用将赋予计算社会学以新的学科能力，助力数据驱动的知识生产，为计算社会学更深入理解和解释复杂社会现象、探索因果关系提供丰富的算法工具箱，也对推动数智时代人文社会科学知识生产范式变革具有深远意义。Abstract: Sociologists have long focused on methods for causal identification and their applications, but have paid relatively little attention to advances in causal discovery. Building on a review of the two theoretical traditions of causal inference, this paper provides a systematic introduction to classical causal discovery algorithms, examining their underlying principles, practical applications, and potential integration with other computational approaches. In an era marked by the rise of large language models, the adoption of causal discovery algorithms offers new disciplinary capabilities for computational sociology, fostering data-driven knowledge production. These algorithms provide a rich toolkit for gaining deeper insights into complex social phenomena and for exploring causal relationships. Moreover, they hold profound implications for advancing a paradigm shift in knowledge production within the humanities and social sciences in the age of digital intelligence.
-
Key words:
- causal inference /
- causal discovery /
- algorithm tools /
- data-driven /
- computational sociology
-
点击查看大图
计量
- 文章访问数: 18
- HTML全文浏览量: 3
下载:
沪公网安备 31010102003103号