CAJ | 학술논문

在应用多层次灰色评价方法的基础上，采用遗传算法优化的神经网络方法（GA－BP神经网络）构建乌鲁木齐市道路交通事故多发路段的鉴别模型。GA－BP神经网络模型预测值与多层次灰色评价结果吻合良好，验证了GA－BP神经网络方法鉴别道路交通事故多发路段的可行性。GA－BP神经网络有着较高的收敛速度和预测精度，为乌鲁木齐市道路交通事故多发路段鉴别提供了一种新的思路与方法。
재응용다층차회색평개방법적기출상，채용유전산법우화적신경망락방법（GA－BP신경망락）구건오로목제시도로교통사고다발로단적감별모형。GA－BP신경망락모형예측치여다층차회색평개결과문합량호，험증료GA－BP신경망락방법감별도로교통사고다발로단적가행성。GA－BP신경망락유착교고적수렴속도화예측정도，위오로목제시도로교통사고다발로단감별제공료일충신적사로여방법。
This article used genetic algorithm to optimize the neural network method (GA-BP neural network)based on the Multi-level Grey Evaluation method,and established the identification model of traffic accident prone sections in Uru-mqi city roads.The predicted values of GA-BP neural network were in good agreement with the multi-level grey evalua-tion results,which verified the feasibility of the identification of roads traffic accident prone sections.GA-BP neural net-work has a higher convergence speed and prediction accuracy,providing a new idea and method for identification of traffic accident prone sections of roads in Urumqi city.