基于随机森林的个体锻炼意愿-行为鸿沟预测变量与组合路径

Predictive Variables and Combined Paths of Physical Activity Intention-Behavior Gap Based on Random Forest Approach

身体活动水平低下是全球共同关注的公共健康问题,实现身体活动促进的挑战之一在于消除锻炼意愿-行为鸿沟。基于对2024年上海市杨浦区的调研数据,建立随机森林模型并进行特征重要度分析,发现:当个体对锻炼行为的感知障碍较小、对缺乏身体活动的感知严重性较大时,出现锻炼意愿-行为鸿沟的概率较低。认为:感知障碍是预测个体锻炼意愿-行为鸿沟存在与否的首要变量,感知严重性以及移动媒介的使用同样有助于消除意愿-行为鸿沟。在此基础上,提出通过降低感知障碍、增强感知严重性,推广移动媒介的使用,进一步开展因果机制研究等策略缓解并消除个体锻炼意愿-行为鸿沟,实现促进身体活动的目标。

The low level of physical activity is a public health issue of common global concern.One of the challenges in promoting physical activity among the population is to eliminate the physical activity intention-behavior gap.By establishing a random forest model and performing feature importance analysis on the survey data from Yangpu District,Shanghai in January 2024,it is found that when an individual's perceived barriers to exercise behavior are smaller and the perceived severity of lack of physical activity is greater,the probability of an exercise intention-behavior gap is lower.It is believed that perceived barriers are the primary variables for predicting the existence or non-existence of an individual's exercise intention-behavior gap,and the perceived severity and the use of mobile media can also help eliminate the intention-behavior gap.On this basis,it is proposed to alleviate and eliminate the individual's exercise intention-behavior gap and achieve physical activity promotion through strategies such as reducing perceived barriers and enhancing perceived severity,promoting the use of mobile media,and furthering causal mechanism research,etc.