统计学在国内医学领域应用趋势的文本挖掘分析

Text Mining Analysis of Trends in the Application of Statistical Methods in China’s Medical Field

摘要: 本文基于CNKI医学博士学位论文，分析国内医学领域常用统计学方法的应用趋势，并结合LDA与ARIMA模型进行主题挖掘与预测。结果显示，ROC曲线分析、Logistic模型、机器学习等方法快速增长，t检验、方差分析等波动下降。LDA提取出模型评估指标、描述性统计、系统性分析、试验设计、统计检验、生存分析、回归分析、多组比较及机器学习九大主题；预测显示系统性分析、统计检验、回归分析及机器学习将持续高热度，描述性统计、多组比较下降。总体呈现“基础稳固、智能与多因素分析加速发展”特征，为医学研究统计学方法选择与规范化应用提供参考。

Abstract: Based on an analysis of medical doctoral dissertations from CNKI, this study examines trends in the application of statistical methods in China’s medical field by integrating LDA and ARIMA models for topic extraction and forecasting. The results show rapid growth in ROC curve analysis, logistic models, and machine learning, while traditional methods such as t-tests and ANOVA exhibit fluctuating declines. LDA identifies nine major topics: model evaluation metrics, descriptive statistics, systematic analysis, experimental design, statistical tests, survival analysis, regression analysis, multiple-group comparisons, and machine learning. Forecasts indicate that systematic analysis, statistical tests, regression analysis, and machine learning will remain highly active, whereas descriptive statistics and multiple-group comparisons are likely to decline. Overall, the field demonstrates “solid foundational methods alongside accelerating development in intelligent and multifactor analyses,” providing valuable guidance for the selection and standardized application of statistical methods in medical research.