Improved Support Vector Machine Models for Work Zone Crash Injury Severity Prediction and Analysis

Mokhtarimousavi, Seyedmirsajad; Anderson, Jason C.; Azizinamini, Atorod; Hadi, Mohammed

doi:10.1177/0361198119845899

Cited by 93 publications

(31 citation statements)

References 55 publications

(64 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…ML methods are more flexible with no or fewer model assumptions for input variables, and also have better fitting characteristics. Some of the commonly used ML approaches used in crash injury severity prediction include artificial neural networks (ANN) [ 58 , 59 , 60 ], random forest [ 54 , 61 , 62 ], support vector machines (SVM) [ 51 , 63 , 64 ], naïve Bayes [ 65 , 66 , 67 ], K-means clustering (KC) [ 68 , 69 , 70 ], and decision trees (DT) [ 71 , 72 , 73 ].…”

Section: Related Workmentioning

confidence: 99%

See 1 more Smart Citation

Exploring the Injury Severity Risk Factors in Fatal Crashes with Neural Network

Jamal

Umer

2020

IJERPH

View full text Add to dashboard Cite

A better understanding of circumstances contributing to the severity outcome of traffic crashes is an important goal of road safety studies. An in-depth crash injury severity analysis is vital for the proactive implementation of appropriate mitigation strategies. This study proposes an improved feed-forward neural network (FFNN) model for predicting injury severity associated with individual crashes using three years (2017–2019) of crash data collected along 15 rural highways in the Kingdom of Saudi Arabia (KSA). A total of 12,566 crashes were recorded during the study period with a binary injury severity outcome (fatal or non-fatal injury) for the variable to be predicted. FFNN architecture with back-propagation (BP) as a training algorithm, logistic as activation function, and six number of hidden neurons in the hidden layer yielded the best model performance. Results of model prediction for the test data were analyzed using different evaluation metrics such as overall accuracy, sensitivity, and specificity. Prediction results showed the adequacy and robust performance of the proposed method. A detailed sensitivity analysis of the optimized NN was also performed to show the impact and relative influence of different predictor variables on resulting crash injury severity. The sensitivity analysis results indicated that factors such as traffic volume, average travel speeds, weather conditions, on-site damage conditions, road and vehicle type, and involvement of pedestrians are the most sensitive variables. The methods applied in this study could be used in big data analysis of crash data, which can serve as a rapid-useful tool for policymakers to improve highway safety.

show abstract

Section: Related Workmentioning

confidence: 99%

“…An ordered probit model was also used as a benchmark to validate the RF model. Mokhtarimousav et al compared the SVM and random parameter mixed logit models for the severity prediction of work zone crashes [ 64 ]. Empirical findings revealed that prediction accuracy from SVM outperformed the mixed logit model.…”

Section: Related Workmentioning

confidence: 99%

Exploring the Injury Severity Risk Factors in Fatal Crashes with Neural Network

Jamal

Umer

2020

IJERPH

View full text Add to dashboard Cite

show abstract

“…The quality assessment comprises a broad set of performance metrics for recognizing the best model. In other similar researches, the Recall is usually accompanied by Precision [22] and F1-Score [23].…”

Section: Purposementioning

confidence: 99%

Long-Term-Based Road Blackspot Screening Procedures by Machine Learning Algorithms

Fiorentini

Losa

2020

Sustainability

View full text Add to dashboard Cite

Screening procedures in road blackspot detection are essential tools for road authorities for quickly gathering insights on the safety level of each road site they manage. This paper suggests a road blackspot screening procedure for two-lane rural roads, relying on five different machine learning algorithms (MLAs) and real long-term traffic data. The network analyzed is the one managed by the Tuscany Region Road Administration, mainly composed of two-lane rural roads. An amount of 995 road sites, where at least one accident occurred in 2012–2016, have been labeled as “Accident Case”. Accordingly, an equal number of sites where no accident occurred in the same period, have been randomly selected and labeled as “Non-Accident Case”. Five different MLAs, namely Logistic Regression, Classification and Regression Tree, Random Forest, K-Nearest Neighbor, and Naïve Bayes, have been trained and validated. The output response of the MLAs, i.e., crash occurrence susceptibility, is a binary categorical variable. Therefore, such algorithms aim to classify a road site as likely safe (“Accident Case”) or potentially susceptible to an accident occurrence (“Non-Accident Case”) over five years. Finally, algorithms have been compared by a set of performance metrics, including precision, recall, F1-score, overall accuracy, confusion matrix, and the Area Under the Receiver Operating Characteristic. Outcomes show that the Random Forest outperforms the other MLAs with an overall accuracy of 73.53%. Furthermore, all the MLAs do not show overfitting issues. Road authorities could consider MLAs to draw up a priority list of on-site inspections and maintenance interventions.

show abstract

“…Apart from the structural merits of ABC implementations, as it said earlier, it also has a considerable impact on roadway safety compared to on-site bridge construction. The effect of work zone presence on crash severity using sophisticated machine learning techniques has been recently investigated by Mokhtarimousavi et al (2019). In their study, different contributing factors that affect injury severity due to work zone presence have been identified and statistically analyzed.…”

Section: Accelerated Bridge Construction (Abc)mentioning

confidence: 99%

Health Monitoring of Closure Joints in Accelerated Bridge Construction: A Review of Non-Destructive Testing Application

Farhangdoust

Mehrabi

2019

ACT

View full text Add to dashboard Cite

Accelerated Bridge Construction (ABC) uses prefabricated elements that are made continuous using cast-in-place joints. Deck joints are normally referred to as "Closure Joints." There have been concerns about long-term durability of these joints that are expected to become rapidly serviceable. Normally, they contain reinforcing bars and enclosures of various shapes that in some cases create congestion within the joint. The specific nature of the joint application, in-situ casting, curing, material incompatibility, cold joints, cavities and steel congestion contribute to creating the potential for leaving defects and anomalies in the closure joints. This, in turn, results in a higher potential for exposure and other detrimental effects with possible degradation in time, and therefore reducing the strength and serviceability of the joint, hence creating a weak link for the structure. The long-term deflections and environmental loading will only exacerbate the situation. Hence, evaluation and health monitoring of the closure joints becomes inevitable. Despite the wide use of non-destructive testing (NDT) methods for bridge structures in general, a concerted attempt for categorization of these methods, comparison of capabilities, and selection of methods most applicable to closure joints is lacking. To address this, a research project was carried out as part of activities in the Accelerated Bridge Construction University Transportation Center (ABC-UTC) of Florida International University. This study included a comprehensive literature review with a focus on NDT methods applicable to health monitoring of ABC closure joints. The study focused on joint types relevant to precast concrete decks commonly used for ABC bridges, therefore, FRP (fiber reinforced plastic), timber (wood), and steel of any shape were excluded for the time being. The study resulted in categorizing the most common closure joints in five general groups based on their features affecting the application of the NDT methods. Accordingly, the most promising NDT methods were identified taking into account the distinctive defects and anomalies associated with closure joints. These methods were evaluated for their efficacy, ease of use and other characteristic influencing their use as preferred methods for each type of joint. A flowchart was introduced to assist in selection of the most applicable NDT method to each type of defect in closure joints. This paper summarizes the results of this study.

show abstract

Improved Support Vector Machine Models for Work Zone Crash Injury Severity Prediction and Analysis

Cited by 93 publications

References 55 publications

Exploring the Injury Severity Risk Factors in Fatal Crashes with Neural Network

Exploring the Injury Severity Risk Factors in Fatal Crashes with Neural Network

Long-Term-Based Road Blackspot Screening Procedures by Machine Learning Algorithms

Health Monitoring of Closure Joints in Accelerated Bridge Construction: A Review of Non-Destructive Testing Application

Contact Info

Product

Resources

About