A Machine Learning-Based Real-time Accident Risk Model for Highway Construction Zones
In recent years,as the machine learning method continuously rises and develops,it has been widely applied in the data mining industry,and it especially performs greatly in data classification and identification.This paper develops a modeling concept for predicting and identifying accident risks.After the preprocessing of raw data,the linear method,the binary Logistic regression,and the non-linear method,the convolutional neural network,are used to construct the accident risk model,respectively.It can determine whether an accident occurs or not according to the traffic flow state in construction zones,resulting in two traffic flow states,the normal passing state,and the high-risk state.This paper further studies and compares the characteristics and modeling effects of the two methods.
machine learningbinary logistic regressionnonlinear methodconvolutional neural networksaccident risk model
万方数据
维普
