JavaScript is disabled for your browser. Some features of this site may not work without it.
| dc.contributor.author | Atsedemariam, Adane Yirdaw | |
| dc.date.accessioned | 2024-11-04T07:39:01Z | |
| dc.date.available | 2024-11-04T07:39:01Z | |
| dc.date.issued | 2023-12 | |
| dc.identifier.uri | http://ir.bdu.edu.et/handle/123456789/16091 | |
| dc.description.abstract | Among the most serious road safety problems is traffic crashes. Road traffic crashes are one of the leading causes of unnatural human deaths worldwide. Due to increasing traffic volume and mobility, road crashes have become a serious problem, especially at nations with low and medium incomes. The developing nations include Ethiopia. . Road traffic accidents are rising alarmingly in Ethiopia, especially in urban areas. An important urban center in Ethiopia is Bahir Dar. Therefore, predicting future traffic crashes is crucial to understanding the problem and speeding up decision-making to address it. Since not every vehicle crash results in the same level of damage and injuries, it is essential to categorize them according to severity. Crash severity analysis requires models for the prediction that are precise and efficient for classifying crashes according to severity. This research aims to develop a model to forecast the severity of crashes using artificial neural networks: Study of a Case of Bahir Dar City. This study evaluated the predictive performance, including prediction accuracy, between an artificial neural network model and a multinomial logistic regression model for crash severity prediction in Bahir Dar City. Models were created using data from 2010 to 2022 by G.C. In this study, the crash day, driver’s gender, driver’s age, educational background of the driver, driving experience, type of vehicle, location, road junction type, weather condition, road condition, road surface, light condition, road type, road feature, cause of the crash, and type of crash were selected and analyzed as model input variables. The hidden and output layers of the ANNs were created using the Adam learning algorithm, as well as the ReLu and Softmax activation functions. For performance evaluation, we used classification accuracy, confusion matrix, and associated evaluation metrics, including accuracy, precision, recall, and F1. In comparison to the best approaches, an artificial neural network method outperformed a multinomial logistic regression model. The accuracy of 65% obtained from the MNL model was clearly less than the ANN accuracy value of 89%. | en_US |
| dc.language.iso | en_US | en_US |
| dc.subject | Civil and Water Resource Engineering | en_US |
| dc.title | Development of a Model for the Prediction of Road Crash Severity Using Artificial Neural Networks: A Case Study of Bahir Dar City | en_US |
| dc.type | Thesis | en_US |
