Using Bayesian Belief Network Techniques and Geographical Information Systems to Predict Highway Crashes in Albania

By Raimonda Dervishi.

Published by Spaces and Flows: An International Journal of Urban and ExtraUrban Studies

Format Price
Article: Print $US10.00
Published online: April 4, 2014 $US5.00

In Albania, a country that is undergoing a tremendous economic and social transformation and modernization, with a rapidly growing web of highways that crisscross the country, no study has been made to predict and prevent traffic crashes which are increasing in number with each passing year. This paper examines research methods used in analyzing highway safety data in various countries as well as proposing a prediction model using Bayesian Belief Network Techniques (BBNT) as part of the efforts to enhance traffic safety data analysis in Albania. We also suggest that Geographical Information Systems (GIS)—which until now have been used primarily for mapping crashes, visual determination for crash occurrence patterns, and identification of high crash spots—can be effectively used in combination with BBNT modelling to enhance the prediction capability for highway crashes, hence the safety patterns as roadway conditions in Albania change.

Keywords: Road Safety, Highway Crashes, Bayesian Methods, Bayesian Belief Network Techniques

Spaces and Flows: An International Journal of Urban and ExtraUrban Studies, Volume 4, Issue 2, June 2014, pp.11-18. Article: Print (Spiral Bound). Published online: April 4, 2014 (Article: Electronic (PDF File; 202.765KB)).

Prof. Raimonda Dervishi

Professor, Department of Mathematics, Politechnical University of Tirana, Tirana, Albania, Albania

Raimonda Dervishi is a doctoral student at the Polytechnic University of Tirana, Albania, where she also teaches applied mathematics and probability.