Data Modeling Approach For Road Accident Prediction: Application To UK Traffic Accident Data

Soumaya Amraoui, Mina Elmaallam, Walid Cherif


The war on the roads continues to wrest the lives of citizens and is very costly to the states worldwide. Several measures have been taken to reduce the number of road accidents. Scientific research remains very active in proposing ideas about new modes of action. The approaches are multiple, but the goal remains the same: reduce the number and severity of accidents.  Designing a model that predicts eventual accidents or predicts the severity of a reported accident is not an easy task, the main problem it faces is the complexity of collected features. In this sense, this paper introduces a new modelling approach that transforms collected features into scaled values, this helps identifying the severity of an accident by applying machine learning techniques. Different machine learning techniques have been compared and decision trees have yielded the highest performance. Afterwards, this work quantifies features’ contributions into the classification phase by using two different techniques, this highlighted most influencing factors. This is of major importance since it proposes some mobilizations for roads accidents prevention and also helps defining the measures to be taken after an accident.


Artificial Neural Network, Data Transformation, Decision Trees, K-Nearest Neighbors, Naïve Bayes, Road Accidents, Support Vector Machines

Total views : 12 times


  • There are currently no refbacks.

Copyright (c) 2019 Institute of Advanced Engineering and Science

Creative Commons License
This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.

View IJAI Stats

Creative Commons License
This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License.