TY - GEN
T1 - Towards Big Data Analytics and Mining for UK Traffic Accident Analysis, Visualization & Prediction
AU - Feng, Mingchen
AU - Zheng, Jiangbin
AU - Ren, Jinchang
AU - Liu, Yanqin
N1 - Publisher Copyright:
© 2020 ACM.
PY - 2020/2/15
Y1 - 2020/2/15
N2 - Road traffic accident (RTA) is a big issue to our society due to it is among the main causes of traffic congestion, human death, health problems, environmental pollution, and economic losses. Facing these fatal and unexpected traffic accidents, understanding what happened and discover factors that relate to them and then make alarms in advance play critical roles for possibly effective traffic management and reduction of accidents. This paper presents our work to establish a novel big data analytics platform for UK traffic accident analysis using machine learning and deep learning techniques. Our system consists of three parts in which we first cluster accident incidents in an interactive Google map to highlight some hotspots and then narratively visualize accident attributes to uncover potentially related factors, finally we explored several state-of-the-art machine learning, deep learning and time series forecasting models to predict the number of road accidents in the future. The experimental results show that our big data processing platform can not only effectively handle large amount of data but also give new insights into what happened and reasonably prediction of what will happen in the future to assist decision making, which will undoubtedly show its great value as a generic platform for other big data analytics fields.
AB - Road traffic accident (RTA) is a big issue to our society due to it is among the main causes of traffic congestion, human death, health problems, environmental pollution, and economic losses. Facing these fatal and unexpected traffic accidents, understanding what happened and discover factors that relate to them and then make alarms in advance play critical roles for possibly effective traffic management and reduction of accidents. This paper presents our work to establish a novel big data analytics platform for UK traffic accident analysis using machine learning and deep learning techniques. Our system consists of three parts in which we first cluster accident incidents in an interactive Google map to highlight some hotspots and then narratively visualize accident attributes to uncover potentially related factors, finally we explored several state-of-the-art machine learning, deep learning and time series forecasting models to predict the number of road accidents in the future. The experimental results show that our big data processing platform can not only effectively handle large amount of data but also give new insights into what happened and reasonably prediction of what will happen in the future to assist decision making, which will undoubtedly show its great value as a generic platform for other big data analytics fields.
KW - Big Data Analytics
KW - Deep Learning
KW - Time series Forecasting
KW - Traffic Accident Analysis
UR - http://www.scopus.com/inward/record.url?scp=85085914127&partnerID=8YFLogxK
U2 - 10.1145/3383972.3384034
DO - 10.1145/3383972.3384034
M3 - 会议稿件
AN - SCOPUS:85085914127
T3 - ACM International Conference Proceeding Series
SP - 225
EP - 229
BT - Proceedings of the 2020 12th International Conference on Machine Learning and Computing, ICMLC 2020
PB - Association for Computing Machinery
T2 - 12th International Conference on Machine Learning and Computing, ICMLC 2020
Y2 - 15 February 2020 through 17 February 2020
ER -
