中国农业大学学报2025，Vol.30Issue(7)：264-274,11.DOI:10.11841/j.issn.1007-4333.2025.07.22
基于机器学习和SHAP算法的我国粮食安全水平测度重构及可解释性分析
Reconstruction and interpretability analysis of China's food security level based on machine learning and SHAP algorithm
摘要
Abstract
To explore the new connotations of China's food security under the concept of the"Big Food Perspective",this study constructs a food security evaluation index system based on the provincial-level data from 2010 to 2022.The framework covers five dimensions,including production security,quality security,economic security,consumption security,and circulation security.Using the entropy method for index calculation and leveraging machine learning and SHAP algorithms,the study re-identifies the importance of various influencing variables on food security.The results reveal that:1)From the perspective of overall evolution trend,China's food security level exhibited a first decreasing and then increasing trend,and there were a rapid growth in 2012-2015 and 2018-2019;2)From the perspective of the evaluation scores of various dimensions,the production security contributed the most to food security,while the economic and quality security showed upward trends.However,the consumption security and circulation security remained the weak links in the food security system;3)From the perspective of factors relatedto food security,the critical determinants of China's food security include total meat production,total grain production,total aquatic product output,railway freight volume,agricultural fiscal expenditure,forest coverage rate,and the urban-rural income gap.Based on the above findings,the study proposed policy recommendations in promoting the sustainable development of animal husbandry,improving grain production efficiency,strengthening aquaculture development,and increasing agricultural fiscal spending.关键词
大食物观/XGBoost/SHAP/粮食安全/食物安全Key words
Big Food Perspective/XGBoost/SHAP/grain security/food security分类
管理科学引用本文复制引用
王火根,胡梦婷,刘小春..基于机器学习和SHAP算法的我国粮食安全水平测度重构及可解释性分析[J].中国农业大学学报,2025,30(7):264-274,11.基金项目
国家自然科学基金项目(71963018) （71963018）
