Improvement of random forest by multiple imputation applied to tower crane accident prediction with missing data

PurposeThis research is aimed at predicting tower crane accident phases with incomplete data.Design/methodology/approachThe tower crane accidents are collected for prediction model training. Random forest (RF) is used to conduct prediction. When there are missing values in the new inputs, they should be filled in advance. Nevertheless, it is difficult to collect complete data on construction site. Thus, the authors use multiple imputation (MI) method to improve RF. Finally the prediction model is applied to a case study.FindingsThe results show that multiple imputation RF (MIRF) can effectively predict tower crane accident when the data are incomplete. This research provides the importance rank of tower crane safety factors. The critical factors should be focused on site, because the missing data affect the prediction results seriously. Also the value of critical factors influences the safety of tower crane.Practical implicationThis research promotes the application of machine learning methods for accident prediction in actual projects. According to the onsite data, the authors can predict the accident phase of tower crane. The results can be used for tower crane accident prevention.Originality/valuePrevious studies have seldom predicted tower crane accidents, especially the phase of accident. This research uses tower crane data collected on site to predict the phase of the tower crane accident. The incomplete data collection is considered in this research according to the actual situation.

Driver behaviour and traffic accident involvement among professional heavy semi-trailer truck drivers in China

The purpose of this study was to develop a driving behavior scale for professional drivers of heavy semi-trailer trucks in China, and study the causes of such driving behavior and its impact on traffic safety operation. Data was processed by IBM SPSS 25. In addition to principal component analysis, Promax rotation, Bartlett’s test, Cronbach’s alpha, correlation analysis and binary logistic regression were examined. A DBQ with 4 dimensions and 20 items, and a PDBQ with 1 dimension and 6 items were developed for professional drivers of heavy semi-trailer trucks in China. The KMO coefficients of PDBQ and DBQ were 0.822 and 0.852, respectively, and the significant level of Bartlett’s popularity test was p < 0.0001. The accident prediction model showed that the variables related to traffic accidents were negligence/lapses and driving time of heavy semi-trailer truck drivers. 1–5 a.m. was found to be the most dangerous period for drivers of medium and heavy semi-trailer trucks, during which accidents were most likely to happen. As negligence/lapses increased by one unit, the probability of traffic accidents increased by 2.293 times.

A Predictive Pedestrian Crash Model Based on Artificial Intelligence Techniques

Every year in Italy, there are about 20,000 road accidents involving pedestrians, with a significant number of injuries and deaths. Out of these, about 30% occur at pedestrian crossings, where pedestrians should be protected the most. Here, we propose a new accident prediction model to improve pedestrian safety assessments that allows us to accurately identify the sites with the largest potential safety improvements and define the best treatments to be applied. The accident prediction model was developed using the ISTAT dataset, including information about the fatal and injurious crashes that occurred in Italy in a 5-year period. The model allowed us to estimate the risk level of a road section through a machine-learning approach. Gradient Boosting seems to be an appropriate tool to fit classification models for its flexibility that allows us to capture non-linear relationships that would be difficult to detect via a classical approach. The results show the ability of the model to perform an accurate analysis of the sites included in the dataset. The locations analyzed have been classified based on the potential risk in the following three classes: High, medium, and low. The proposed model represents a solid and reliable tool for practitioners to perform accident analysis with pedestrian involvement.

Review and analysis of roadway crash prediction studies on urban roads under heterogeneous traffic conditions

In recent decades, pre-predicting the roadway accidents is essential for real time traffic incident management that effectively minimizes the environmental pollution, traffic congestion and secondary incidents. Currently, the traffic data are available in thousands of public and private datasets and also generates terabytes of data each year. Though, it is infeasible to manage the huge datasets by utilizing traditional software and hardware. It is therefore essential that an automated system to predict road accidents is developed. The present review paper investigates the researches done on road accident prediction, particularly for urban roads under heterogeneous traffic conditions. It also explores the problems faced in existing works by researchers. This review paper helps researchers achieve a better solution for the current problems faced by heterogeneous traffic conditions when it comes to urban road accident prediction. The findings demonstrate that the operating speed and the disparities between the speed restrictions and the operating speed are the key factors influencing the accident frequency rate.

Statistical Analysis and Development of Accident Prediction Model of Road Safety Conditions in Hisar City

The severity of road accidents is a big problem around the world, particularly in developing countries. Recognizing the major contributing variables can help reduce the severity of traffic accidents. This research uncovered new information as well as the most substantial target-specific factors related to the severity of road accidents. T-stat, P-value, Significance and other test values are determined to check the dependency of dependent variable on independent variable in order to obtain the most significant road accident variables. In this research, a comparative analysis of accident data from Hisar and Haryana are compared. According to the findings, Haryana’s accident severity index (46.20) was higher in 2019 than Hisar’s (36.01), while Hisar had fewer accidents per lakh population (33.34) than Haryana (38.40). The outcomes of the study were used to develop an effective and precise accident predicting model is developed for Hisar city and state Haryana using a statistical method. Four models were created using linear regression analysis, two each for Hisar and Haryana. These models produce good results with a margin of error that is within acceptable bounds (0-5%), allowing them to be used to predict future traffic accidents and deaths.