驾驶员行为模式的因子分析和模糊聚类

为了研究驾驶员的驾驶行为模式及人群聚类特点, 基于标准化驾驶员行为问卷开展了以中国非职业驾驶员为样本的驾驶行为自填式调查, 通过统计分析验证了问卷的有效性和可信性, 并基于因子分析方法, 从问卷结果中提取出驾驶能力因子和驾驶风格因子, 实现了对驾驶员驾驶行为模式的量化描述。进而以提取出的因子为特征, 采用模糊C均值聚类分析的方法建立了驾驶员行为模式的类别分布空间。
对比分析表明: 驾驶能力因子值越大, 驾驶员驾驶能力越低; 驾驶风格因子值越大, 驾驶员驾驶风格越激进; 聚类结果与驾驶员主观自我评估的分布趋势一致, 能力偏弱、风格偏激进的聚类人群具有较高的事故驾驶员比率。

Factor analysis and fuzzy clustering of driver behavior pattern

In order to study driver pattern and population clustering feature, driver behavior self-reported investigation for normal drivers was conducted with standardized driver behavior questionnaire(DBQ), and its validity and reliability of the questionnaire were verified based on statistics analysis.Driving ability factor and driving style factor were extracted from the questionnaire results to quantify driver characteristics with factor analysis method.Based on fuzzy C-means clustering algorithm, driver cluster distribution was established with the two factors as pattern features.
Comparison result indicates that the higher the driving ability factor value is, the weaker the driving ability is; the higher the driving style factor value is, the more aggressive the driver style is.The clustering result has similar distribution tendency with driver subjective evaluation, and the clusters with weaker ability or more aggressive style have higher accident driver rate.