Intelligent Data Analysis

11Applying Big Data Analytics on Motor Vehicle Collision Predictions in New York City

Dhanushka Abeyratne and Malka N. Halgamuge

Department of Electrical and Electronic Engineering, Yellowfin (HQ), The University of Melbourne, Australia

11.1 Introduction

11.1.1 Overview of Big Data Analytics on Motor Vehicle Collision Predictions

Due to population growth, there are more traffic accidents, which have become a global concern. In the past, researchers have conducted studies to find out the common cause for motor vehicle collisions. Log-linear models, data mining, logical formulation, and fuzzy ART are some of the methods widely used to perform research [1]. Even with the use of these methods, data analysis is a complicated process. However, with the improvements in technology, data mining is defined to be a highly accurate method for the analysis of big data.

With the use of big data application, few researchers have focused mainly on understanding the significance of vehicle collisions. A research performed by Shi et al. [2] explains the significance of identifying the traffic flow movements on highways to minimize the impact on vehicle collisions. This research has used a time series data approach under clustering analysis to comprehend traffic flow movements. It was converted using the cell transformation method.

A similar study by Yu et al. [3] considers traffic data mining to be a big data approach. The author further carries out the study using past traffic data with the ...