基于文本挖掘的我国老龄化主题热点识别研究

Research on Hotspot Identification of China’s Aging Topics Based on Text Mining

摘要: 2021~2025年是我国积极应对人口老龄化国家战略全面落地的关键时期。本文基于中国知网11,758篇核心期刊文献，采用BERTopic模型提取老龄化研究主题并识别热点。研究表明：30个细分主题可归为四大集群——“养老服务与政策保障”(47.1%)为核心主线，“老年健康与慢病管理”(32.7%)为基础支撑，“数字适老与智慧养老”(11.8%)为前沿方向，“银发经济与社会支持”(8.4%)为交叉增长域。集群间关联分析显示，政策集群与技术集群呈强协同关系(相似度0.65~0.72)，而老年医学类主题与社科管理类主题之间关联较弱(相似度 < 0.4)，提示存在学科壁垒。农村老龄化(0.8%)与人工智能伦理(0.5%)等主题文献占比较低，相关领域亟待深化。本研究为老龄化研究的选题规划与政策决策提供了基于大数据的实证参考。

Abstract: The period from 2021 to 2025 represents a critical phase for the full implementation of China’s national strategy to actively respond to population aging. Based on 11,758 core journal articles indexed in CNKI, this study employs the BERTopic model to extract research topics and identify hotspots in the field of aging. The findings reveal that 30 refined topics can be grouped into four major clusters: “Elderly Care Services and Policy Guarantees” (47.1%) as the core thread, “Elderly Health and Chronic Disease Management” (32.7%) as the foundational support, “Digital Age-Friendliness and Smart Elderly Care” (11.8%) as the frontier direction, and “Silver Economy and Social Support” (8.4%) as a cross-cutting growth domain. Inter-cluster association analysis indicates a strong synergistic relationship between the policy cluster and the technology cluster (similarity: 0.65~0.72), while the association between geriatric medicine-related topics and social science/management-related topics is relatively weak (similarity < 0.4), suggesting the existence of disciplinary barriers. Topics such as rural aging (0.8%) and AI ethics (0.5%) account for relatively low proportions of the literature, indicating areas in urgent need of further exploration. This study provides empirical, big-data-based evidence for topic selection and policy decision-making in aging research.