Big Data Analysis for Predicting Road Traffic Accident Severity

B.Sc. Ametu Negash Redi

Master's thesis

Abstract:

Abstract With the recent growth of urban transportation, Frequent road traffic accident and other issues must be addressed. Understanding the causes of traffic accidents and developing an early alarm model for drivers and pedestrians would be critical in resolving traffic accident issues. To achieve the end goal and scopes, this study relies on Big Data as a primary source. It exhibits typical characteristics and issues that are arisen when dealing with Big Data. The focus of the thesis is to develop an in-depth analysis of traffic accident severity and to determine which of the predictive models used is the best one for predicting the occurrence and the level of accident severity as the target variable. The three models are selected for the practical part i.e., Logistic Regression, Decision tree, and Neural network and to identify the best predictive model, models are tested using different performance measurement parameters Root mean square error (RMSE), receiver operating characteristics (ROC). Based on the model comparison in SAS model studio, the test results show that the decision tree seemed to perform better than logistic regression and neural network models. This research study shows that the algorithm can predict accidents with 87% accuracy. The dataset was taken from the UK traffic accident from 2005 to 2015 retrieved from Kaggle source. The aim of analyzing this dataset was for academic purposes only, and the data would not be used for any other purpose. It includes 1,048,575 records of accidents with various information about a traffic accident.

Keywords

Traffic accident, Big data, predictive modeling, Decision tree, Logistic Regression, Neural network.

Jazyk práce: English

Datum vytvoření / odevzdání či podání práce: 31. 3. 2021

Obhajoba závěrečné práce

• Vedúci: Ing. Tomáš Hlavsa, Ph.D.

Czech University of Life Sciences Prague

Faculty of Economics and Management

Master programme / odbor:
Systems Engineering and Informatics / Informatics