习惯性违章行为耦合关联分析模型

Couplinganalysis model for habitual violation behavior

为探讨习惯性违章行为(HVB)的属性间与属性内的关系特征,建立习惯性违章行为的耦合关联分析模型。首先,分析违章行为属性值的分布特征及关联关系,运用关联规则(ARM)挖掘思想和耦合关系理论对各类违章行为下相应属性的关联系数进行求解,得到一个耦合关联度向量集,且对诸向量从大到小排序;然后,依据排序后耦合关联度向量集映射成习惯性违章行为耦合关联分析模型。最后,引入召回率、精确率和平均绝对值误差(MAE)等3个指标,分别求解数据集和模型的指标结果。所建模型与ARM分析结果的对比表明,模型在习惯性违章行为关联关系分析的准确性与全面性方面都效果良好。

In order to explore the relational characteristics between and within the attributes of HVB,a coupling association analysis model of HVB should be built. Firstly,distribution characteristics and association relationships of the violation attribute values were analyzed. The correlation coefficients of the corresponding attributes under various violations were obtained by using the ARM and coupling relation theory.A set of coupling correlation degree vectors,meanwhile,was obtained,and the vectors were ranked from large to small. And then,according to the sorted coupling correlation degree vector set,the coupling association analysis model of HVB was mapped. In the end,3 indicators,recall rate,precision rate and mean absolute error（ MAE） were introduced to solve the indicators results of the data set and the model. A comparison was made between the model and ARM analysis results. The result shows that the model performs well in accuracy and comprehensiveness for habitual violation behavior＇s association relationship analysis.