This study applied GIS-based statistical analytic techniques to investigate the influence of accident Severity Index (SI) on temporal-spatial patterns of accident hotspots related to the specific time intervals of day and seasons. Road Traffic Accident (RTA) data in 3 years (2015 − 2017) in Hanoi, Vietnam were used to analyze and test this approach. Firstly, the RTA data were divided into four seasons in accordance with Hanoi’s weather conditions and the time intervals such as the daytime, nighttime, or peak hours. Then, the Kernel Density Estimation (KDE) method was applied to analyze hotspots according to the time intervals and seasons. Finally, the results were presented by using the comap technique. This study considered both analyses with and without SI. The accident SI measures the seriousness of an accident. The approach method is to give higher weights to the more serious accidents, but not with the extremely high values calculated on a direct rate to the accident expenditures. The results showed that both analyses determined the relatively similar hotspots, but the rankings of some hotspots were quite different due to the integration of SI. It is better to take into account SI in determining RTA hotspots because the gained results are more precise and the rankings of hotspots are more accurate. From there, the traffic authorities can easily understand the causes behind each accident and provide reasonable solutions to solve the most dangerous hotspots in case of limited budget and resources appropriately. This is also the first study about this issue in Vietnam, so the contribution of the article will help the traffic authorities easily solve this problem not only in Hanoi but also in other cities.

The study presented an approach that integrated kernel density estimation (KDE) algorithm and spatial autocorrelation analysis, which helped to determine traffic accident (TA) hotspot locations and simultaneously evaluate the statistical significance of the hotspot clusters. Firstly, hotspots were identified by applying a GIS-based KDE algorithm. Secondly, the hotspot clusters were evaluated in terms of statistical significance by applying the Moran’s I statistic indices. Finally, hotspots were arranged according to their significance. TA data in the (2015–2017) period in Hanoi, Vietnam were applied to test this approach. Importantly, the study proposed a validation process for the results by applying the Gi* statistics method to validate the (H-H) clusters and the EPDO method to validate the ranking of hotspots. The results showed that this integration overcame the drawbacks of the KDE method. The statistical test process of clusters helped to hinder the occurrence of too many clusters that were determined by the KDE method because they were not really dangerous. This approach was helpful and precise in identifying TA hotspots with the statistical meaning. These outcomes will not only enable traffic authorities to comprehensively understand the reasons for each accident but also to help them manage and deal with hazardous areas according to the prior order in case of limited expenses and allocate traffic safety sources accordingly.

Urban traffic accidents pose significant challenges to the sustainability of transportation infrastructure not only in Vietnam but also all over the world. To decrease the frequency of accidents, it is crucial to analyze accident data to determine the relationship between accidents and causes, especially for serious accidents. This study suggests an integrated approach using Geographic Information System (GIS) and Data Mining methods to investigate the features of urban traffic accidents in Hanoi, Vietnam aiming to solve these challenges and enhance the safety and efficiency of urban transportation. Firstly, the dataset was segmented into homogenous clusters using the two-step cluster method. Secondly, the correlation between causes and traffic accidents was examined on the overall dataset as well as on each cluster using the association rule mining (ARM) technique. Finally, the location of accident groups and high-frequency sites of accidents (hotspots) were determined by using GIS techniques. As a result, a five-cluster model was created, which corresponded to five common accident groupings in Hanoi. Moreover, the results of the study also identified the types of accidents, the main causes, the time, and the surrounding areas corresponding to each accident group. In detail, cluster 5 depicted accidents on streets, provincial, and national roads caused by motorbikes making up the highest percentage within the groups, accounting for 29.2%. Speeding and driving in the wrong lane in the afternoon and at night were the main causes in this cluster ($C_f$ ≥ 0.9 and $L_t$ ≥ 1.22). Next, cluster 2 had the second-highest proportion. Cluster 2 presented accidents between a truck/car and a motorbike on national and provincial roads, accounting for 27.8%. Cluster 1 presented accidents between a truck/car and a motorbike on local streets, accounting for 22%. Cluster 3 illustrated accidents between two motorbikes on the country lanes, accounting for 12.3%. Finally, cluster 4 depicted single-vehicle motorbike crashes, with the lowest rate of 8.8%. More importantly, this study also recommended using repeatability criteria for the same type of accidents or causes to determine the location of hotspots. Also, suggestions for improving traffic infrastructure sustainability were proposed. To our knowledge, this is the first time in which these three methods are applied simultaneously for analyzing traffic accidents.

Cluster analysis is often employed as the initial stage in organizing heterogeneous data into homogeneous groups. Choosing an effective clustering approach and an ideal number of clusters in a traffic accident dataset might be complex and challenging. This study aims to evaluate the effectiveness of k-means and two-step cluster methods. Subsequently, the two-step cluster method and GIS are applied to analyze the traffic accident datasets from 2015 to 2017 in Hanoi, Vietnam. First, according to the Silhouette score, the two-step cluster method achieved a higher score of 0.563, while the k-means method scored 0.341. A higher Silhouette score indicates more well-defined clusters. Second, the research suggests combining the two-step cluster method with GIS for analyzing traffic accident datasets. The outcome identifies five typical types of accidents in Hanoi. In addition, the locations of various accident types were visually illustrated on a map, enabling traffic officials to recommend precise and urgent countermeasures. Importantly, the clustering results reveal that the two-step cluster method exhibits a significantly higher rate of homogeneous data in the clusters compared to the k-means method. This study demonstrates that the two-step cluster method is not only more effective than the k-means method in terms of clustering ability but also in data pre-processing. The study's results enable authorities to gain a more detailed understanding of typical traffic accident patterns in Hanoi. Besides, the employed methods could potentially be applied to other regions, providing an additional avenue for analysis

Early forecasting of vehicle flow speeds is crucial for sustainable traffic development and establishing Traffic Speed Forecasting (TSF) systems for each country. While online mapping services offer significant benefits, dependence on them hampers the development of domestic alternative platforms, impeding sustainable traffic management and posing security risks. There is an urgent need for research to explore sustainable solutions, such as leveraging Global Positioning System (GPS) probe data, to support transportation management in urban areas effectively. Despite their vast potential, GPS probe data often present challenges, particularly in urban areas, including interference signals and missing data. This paper addresses these challenges by proposing a process for handling anomalous and missing GPS signals from probe vehicles on parallel multilane roads in Vietnam. Additionally, the paper investigates the effectiveness of techniques such as Particle Swarm Optimization Long Short-Term Memory (PSO-LSTM) and Genetic Algorithm Long Short-Term Memory (GA-LSTM) in enhancing LSTM networks for TSF using GPS data. Through empirical analysis, this paper demonstrates the efficacy of PSO-LSTM and GA-LSTM compared to existing methods and the state-of-the-art LSTM approach. Performance metrics such as Root Mean Square Error (RMSE), Mean Absolute Error (MAE), and Median Absolute Error (MDAE) validate the proposed models, providing insights into their forecasting accuracy. The paper also offers a comprehensive process for handling GPS outlier data and applying GA and PSO algorithms to enhance LSTM network quality in TSF, enabling researchers to streamline calculations and improve supposed model efficiency in similar contexts.

Road traffic crashes pose a serious public safety challenge, particularly due to fatal and serious injuries. Although machine learning (ML) has been widely used for crash severity prediction, many studies remain accuracy-oriented and insufficiently address class imbalance, decision thresholds, and probabilistic reliability. This study proposes a safety-oriented and explainable ML framework for predicting killed or seriously injured (KSI) crashes using nationwide United Kingdom traffic accident data from 2020–2024. Crash severity is reformulated as a binary classification task distinguishing slight injury crashes from KSI outcomes, aligning model objectives with road safety priorities. A Light Gradient Boosting Machine (LightGBM) model is developed with imbalance handling using SMOTE, safety-oriented decision threshold optimization, and probability calibration. Model performance is evaluated using ROC–AUC, precision–recall analysis, confusion matrices, the Brier score, and a utility-based evaluation metric, while interpretability is ensured through SHapley Additive exPlanations (SHAP). Results show that default threshold settings fail to adequately detect severe crashes. At an optimized threshold of 0.35, the model achieves a Recall(KSI) of 0.605 – representing a substantial 73% improvement compared to conventional configurations – while maintaining acceptable precision. In addition, probability calibration confirms reliable risk estimation (Brier score = 0.190), supporting risk-based interpretation. Comparative analysis demonstrates that the SMOTE-based model provides a more balanced and operationally effective trade-off than class-weighted learning. SHAP analysis identifies speed limit, road class, lighting conditions, and urban context as key variables associated with KSI risk. The findings highlight the importance of safety-oriented learning design and context-aware performance interpretation for effective, risk-based traffic safety management.