Paper Info
Title
Finding Factors and Vehicles Involved in Two-Vehicle Accidents Through the Use of Social Network Analysis.
Abstract
Social Network Analysis (SNA) has emerged as a new paradigm to effectively represent complex patterns of relationships between all categories of social groups. It helps to find the structure of ties and its impact on individuals, groups or even incidents. This article is a similar attempt to explore the vehicles and vehicle-related violations leading to accidents through the use of SNA. SNA is performed on accident data of New York for 2016. SNA measures including degree centrality, betweenness centrality and eigenvector centrality are used to probe into the impact of ties between actors of accidents. The empirical analysis shows that 'Passenger vehicle' with degree centrality of 7513 has the highest degree of accidents, explaining 41.12% of total accidents. In addition, it is involved in 57.42% of accidents when accidents occur between same types of vehicles. 'Sport utility/station wagon' and 'taxi' rank second and third in this category with degree values of 4657 and 1454 respectively. It is also found that 'driver inattention' holds the pivotal place when violations leading to accidents are concerned. It accounts for 19.09% accidents in general and 44.58% when accidents, where both parties commit the same violation, are considered. 'Failure to yield right-of-way' and 'following too closely' are ranked second and third. Research also finds that Manhattan area of New York is marked by elevated number of accidents.
Year
DOI
Venue
2017
10.1145/3110025.3116210
ASONAM '17: Advances in Social Networks Analysis and Mining 2017 Sydney Australia July, 2017
Keywords
Field
DocType
social network analysis,accidents,degree centrality,eigenvector centrality,traffic violations,accident fatalities
Social psychology,Social group,Ranking,Eigenvector centrality,Commit,Computer science,Computer security,Social network analysis,Centrality,Betweenness centrality,Artificial intelligence,Machine learning
Conference
ISSN
ISBN
Citations 
2473-9928
978-1-4503-4993-2
0
PageRank 
References 
Authors
0.34
1
3
Name
Order
Citations
PageRank
Imran Ashraf137330.26
Soojung Hur2217.34
Yongwan Park322527.97