Traffic accident duration prediction using text mining and ensemble learning on expressways

Abstract

Predicting traffic accident duration is necessary for ensuring traffic safety. Several attempts have been made to achieve high prediction accuracy, but researchers have not considered traffic accident text data in much detail. The limited text data of the first report on an incident describes the characteristics of an accident that are initially available. This paper uses text data fusing and ensemble learning algorithms to build a model to predict an accident’s duration, and a preprocessing scheme of accident duration text data is established. Next, the random forest (RF) algorithm is applied to select feature variables of text data related to the traffic incident duration. Last, a text feature vector is introduced to models such as decision tree, k nearest neighbor, support vector regression, random forest, Gradient Boosting Decision Tree, and Xtreme Gradient Boosting. Our results show that the improved RF model has good prediction accuracy with RMSE, MAPE and R2. From this, the textual factors important to determining the duration of the accident are identified. Further, we investigated that the cumulative importance of 60% is sufficient for traffic accident prediction using text data. These results provide insights into minimizing traffic congestion related to accidents and contribute to the input optimization in text prediction.

Traffic accident duration prediction using multi-mode data and ensemble deep learning

“…The benefits of the various feature variables considered on accident duration predictions were evaluated by comparing the weighted F1 scores obtained using the various modes in conjunction with various prediction models. We evaluated the benefits of the conventional feature variables in Modes 1 and 2 using various conventional and more recent methods, including decision tree [ 14 ], random forest [ 11 ], adaboost [ 19 ], gradient-boosted decision tree (GBDT) [ 34 ], catboost [ 19 ], extra tree [ 19 ], XGBoost [ 19 ], and the proposed BiGRU [ 36 ] and BiGRU-CNN [ 36 ] deep learning methods. In addition, the benefits of the feature variables in Modes 3 and 4 were evaluated using deep-learning-based BiLSTM [ 37 ], LSTM-CNN, BiLSTM-CNN [ 34 ], GRU-CNN, BiGRU, and BiGRU-CNN [ 36 ] methods.…”

“…We evaluated the benefits of the conventional feature variables in Modes 1 and 2 using various conventional and more recent methods, including decision tree [ 14 ], random forest [ 11 ], adaboost [ 19 ], gradient-boosted decision tree (GBDT) [ 34 ], catboost [ 19 ], extra tree [ 19 ], XGBoost [ 19 ], and the proposed BiGRU [ 36 ] and BiGRU-CNN [ 36 ] deep learning methods. In addition, the benefits of the feature variables in Modes 3 and 4 were evaluated using deep-learning-based BiLSTM [ 37 ], LSTM-CNN, BiLSTM-CNN [ 34 ], GRU-CNN, BiGRU, and BiGRU-CNN [ 36 ] methods. These methods were applied because hybrid deep learning networks based on LSTM and GRU have been demonstrated to be suitable for use with text data, while conventional machine learning methods are not effective in processing high dimensional sparse text vectors.…”

“…Finally, Chen et al applied text data fusion and ensemble learning algorithms to predict accident duration based on specific feature variables of text data selected by an RF algorithm. The results show that the MAPE value is stable within 73–75% [ 34 ].…”

Integrative Approach for High-Speed Road Surface Monitoring: A Convergence of Robotics, Edge Computing, and Advanced Object Detection

“…The system employs YOLOv3 and Kalman filters for vehicle detection and tracking, providing auxiliary research for road accidents [9]. The rapid response to highway accidents, the timely arrival of emergency personnel, effective accident scene management, and casualty care are also subjects of extensive research by many scholars [10]. Zou and colleagues conducted a visual exploration of the knowledge base, thematic distribution, research frontiers, and trends in road safety research.…”

Research on the prediction of traffic accident by linear regression

“…In addition, traffic accident creates congestion, causing the transportation system to malfunction. Then the emission of waste gas intensifies, which will further ruin the air [3]. In general, whether it causes human casualties or environmental health, it will cause an inefficient use of social resources.…”