Application of Natural Language Processing Methods in Exploring Changes in Parent-Child Relationships under the "Double Reduction" Policy and Their Key Predictors

This study employs natural language processing (NLP) methods to explore the impact of Chinas "Double Reduction" policy on parent-child relationships and identify key predictive factors. By quantitatively analyzing 1,433 textual data samples, the study applies NLP algorithms, utilizing the Word2Vec pre-trained model and introducing the Dynamic Topic Model (DTM) to examine how parental roles, educational approaches, education policies, and other key factors influence parent-child relationships. The results indicate that: (1) In the initial phase of policy implementation, parents focused more on balancing work and family life, gradually shifting towards emphasizing their children's holistic development and social competence over time. (2) During policy adjustments, parental cognitive changes and responsibility anxiety played an intermediary role in adjusting their roles and interactions with their childrenthat is, changes in cognition and responsibility anxiety prompted shifts in parental roles. This study not only provides new perspectives and methodological support for research in psychology and education but also offers scientific evidence for optimizing family education environments and policymaking.