Predicting the severity of road traffic accidents in the city of Barcelona
Tipus de documentTreball Final de Grau
Data2021-05-14
Condicions d'accésAccés obert
Llevat que s'hi indiqui el contrari, els
continguts d'aquesta obra estan subjectes a la llicència de Creative Commons
:
Reconeixement-NoComercial-CompartirIgual 3.0 Espanya
Abstract
The study of road traffic accidents and the adoption of measures to reduce them is one of the most important targets of the Sustainable Development Goals for 2030. In one of the major cities in EU, Barcelona, 205 people were badly injured with 30 of them having a fatal outcome and more than 10000 were mildly injured in 2015. In this project, accidents in Barcelona comprised between 2018 and 2019 are studied. It focuses on using machine learning techniques to analyze driver-injury severity. Predicting accident-related injury severity provides insights that benefit the prevention of casualties, which supposes a huge economic and social impact. That is, classifying existing casualties into different severity categories allows to explain the underlying patterns behind casualties. A random forest (RF) and classification tree(CART) approach is followed. The former is used to select the most important variables, which are then used as input for the latter; in this way, the variance is reduced and the classification rate is improved. Explanatory variables regarding driver, accident and vehicle characteristics, environmental status, and traffic density are used as input. It is worth to mention the theoretical aspect of this project, the aim is to explain the causes of lower and higher injury severity in existing casualties rather than proposing a methodology to decrease the severity. Also, since multiple types of accidents and vehicles are taken into account the scope of the results is wider. Studies focused on single types and methodologies to prevent severe outcomes can be born from this project
MatèriesAutomatic classification, Traffic safety -- Evaluation -- Statistical methods, Traffic safety -- Spain -- Barcelona, Barcelona Metropolitan Area (Spain) -- Statistics, Traffic accidents -- Statistics, Seguretat viària -- Catalunya -- Barcelona, Seguretat viària -- Avaluació -- Mètodes estadístics, Classificació automàtica, Barcelona (Catalunya : Àrea Metropolitana) -- Estadístiques, Accidents de tràfic -- Estadístiques
TitulacióGRAU EN ENGINYERIA EN TECNOLOGIES INDUSTRIALS (Pla 2010)
Localització
Fitxers | Descripció | Mida | Format | Visualitza |
---|---|---|---|---|
mem-ria.pdf | 1,767Mb | Visualitza/Obre | ||
annex.pdf | 2,015Mb | Visualitza/Obre |