Abstract: The aging process in China is accelerating. In 2020, the Ministry of Public Security has announced the abolition of the upper limit of 70 years of age for small car drivers license applications. The driving behavior of elderly people has attracted much more attention. A comprehensive analysis method combining the structural equation model and Bayesian network was proposed. The structural equation model was used to verify the social cognitive theoretical factors of elderly peoples switching intention toward driving, Bayesian networks in the field of machine learning was combined to calculate and output key determinants and the relationship strength among influencing factors were also quantified. The research results show that about 16% of the elderly have high-level switching intention toward driving, and attitude, subjective norms, and self-efficacy have the greatest impact on switching intention toward driving. When self-efficacy and attitude of the elderly people change from a low state to a high state, the probability of occurrence of a high state of switching intention increases by approximately 185% and 265%, which can help to understand the psychological needs of elderly drivers and provides ideas for predicting the size of the elderly driving group in the future.

Key words: traffic and transportation engineering; driving test policy; elderly people; switching intention; social cognitive theory; Bayesian network

摘要： 我国老龄化进程不断加快，公安部于2020年发布取消小型汽车驾驶证申请70周岁上限，老年人群的驾车出行行为备受关注。提出结构方程模型与贝叶斯网络的综合分析方法：以结构方程模型验证老年人转向驾车出行意向的社会认知理论因素，结合机器学习领域的贝叶斯网络计算输出关键决定因素并量化因素间的关系强度。研究结果表明：约16%的老年人有高水平的驾车出行转移意向；态度、主观规范与自我效能感对驾车出行转移意向产生的影响力最大；当老年人的自我效能感和态度由低状态变为高状态时，转移意向的高状态发生概率呈现约185%和265%的增长幅度。这有助于了解老年人的驾车心理需求，为预测未来高龄驾驶群体规模提供思路。

关键词: 交通运输工程；驾考政策；老年人；转移意向；社会认知理论；贝叶斯网络