基于机器学习的国家认同的影响因素研究
Research on the Influencing Factors of National Identity Based on Machine Learning

本文基于2019年中国社会状况综合调查(CSS2019)，探索国家认同的影响因素。使用数据清洗后的266个数据变量建立MLP模型，来预测被试的国家认同，使用DALEX包来探索最具价值的预测因子。总体来说，机器学习模型确定的预测因子与现有的研究结论是一致的，例如：经济因素、宗教、社会保障等。同时，本文补充了前人对于国家认同的影响因素的研究，为“对法治的感受”可影响国家认同提供了数据支撑。我国应全面推进依法治国，加快建设社会主义法治国家。同时，本文也是将机器学习技术应用于心理学领域的一次积极尝试。
Based on the 2019 Chinese Social Survey (CSS 2019), this research explores the influencing factors of national identity. After data preprocessing, we use the 266 variables to build an MLP model to predict the national identity of the subjects and use the DALEX package to select the most valuable predictor. In general, the predictive factors determined by the machine learning model are con-sistent with results of prior studies, such as economic factors, religion, social security, etc. This re-search supplements previous studies on the factors affecting national identity, and provides data support for “feelings of the rule of law” that can affect national identity. China should comprehen-sively promote the rule of law and accelerate the construction of a socialist country ruled by law. At the same time, this research is an active attempt to apply machine learning technology to the field of psychology.