As the limitations of the single classifier on traffic accident severity, this paper proposed an ensemble approach for improving the prediction performance. It used CNN(convolutional neural network) to extract the features from the spatial dimension, got an ensemble approach with XGBoost and CNN by stacking(multi-level boosting algorithm). The predicting precision of the approach is 91.51% on the validation set. In comparison with the single classification model, the result of the experiment shows a better performance. For providing useful information for reducing the number of traffic accidents and downgrading the severity of traffic accident, the paper gave out a correlation analysis by sorting the features based on the predictions.

[2019-08-05] Printed Article /doc/2018.03.0273/v2