Title: Call detail record-based traffic density analysis using global K-means clustering

Authors: Suja Chandrasekharan Nair; M. Sudheep Elayidom; Sasi Gopalan

Addresses: School of Engineering, Cochin University of Science and Technology, Cochin 682021, India ' School of Engineering, Cochin University of Science and Technology, Cochin 682021, India ' School of Engineering, Cochin University of Science and Technology, Cochin 682021, India

Abstract: With the expanding number of vehicles on the road is creating substantial traffic that is hard to control and maintain safety, particularly in extensive urban areas. To estimate the traffic density several works were carried out in the past. However, they are inappropriate and expensive due to the dynamics of traffic flow. Here we intend to use CDR to distinguish the traffic density location and to track the location of the mobile user. In our proposed method to discover the density scope of the traffic, we are using two algorithms called k-means clustering and the k nearest neighbour classification algorithms. The proposed technique will be tested among five different locations during the weekdays and the weekends, which show the noteworthiness of the proposed algorithm and show that our technique has high accuracy.

Keywords: traffic density; call detail records; CDR; data pre-processing; global K-means clustering algorithm; K-nearest neighbour classification; cell-tower ID; behavioral patterns; disposition; monitoring; predictable.

DOI: 10.1504/IJIE.2020.104654

International Journal of Intelligent Enterprise, 2020 Vol.7 No.1/2/3, pp.176 - 187

Received: 30 Sep 2018
Accepted: 14 Jan 2019

Published online: 24 Jan 2020 *

Full-text access for editors Access for subscribers Purchase this article Comment on this article