A novel approach for traffic accident analysis and prediction using Artificial Neural Networks

S. M. Al-Alawi, G. A. Ali, C. S. Bakheit

Research output: Contribution to journalArticle

7 Citations (Scopus)

Abstract

Traffic accidents are among the major causes of death in the Sultanate of Oman, especially in the age group of 16 to 29. The fatality rate per 10, 000 vehicles appears to be one of the highest in the world Artificial Neural Network (ANN) is a powerful technique that has demonstrated remarkable success in the analysis of historical data and in predicting future trends in many engineering fields. This novel technique was used to analyse the number of car accidents in the Sultanate of Oman during the period from 1976 to 1990. Input for the model was carefully selected through examining the strength of the correlation between the number of accidents and several related variables such as population growth, gross domestic product, number of vehicles on the road, etc. Results indicate that 95.5% of the variation in the number of accidents could be explained by the model. Predictions for the years 1991-1994 showed high accuracy (92%). To further validate the model, principal component analysis (PCA) regression technique was used to fit the same data, and predictions for 1991-1994 were obtained. Statistical analysis of the results showed that the ANN model prediction gave a lower mean absolute percentage error (MAPE) compared to the PCA. In addition, it also gave a lower MSE and a higher R2 value. These results demonstrate that ANN provided closer predictions to the actual results than PCA.

Original languageEnglish
Pages (from-to)118-127
Number of pages10
JournalRoad and Transport Research
Volume5
Issue number2
Publication statusPublished - Jun 1996

Fingerprint

Highway accidents
Principal component analysis
Neural networks
Accidents
Statistical methods
Railroad cars

ASJC Scopus subject areas

  • Civil and Structural Engineering
  • Mechanical Engineering

Cite this

A novel approach for traffic accident analysis and prediction using Artificial Neural Networks. / Al-Alawi, S. M.; Ali, G. A.; Bakheit, C. S.

In: Road and Transport Research, Vol. 5, No. 2, 06.1996, p. 118-127.

Research output: Contribution to journalArticle

@article{4295021dd2eb44258857b88bc1007651,
title = "A novel approach for traffic accident analysis and prediction using Artificial Neural Networks",
abstract = "Traffic accidents are among the major causes of death in the Sultanate of Oman, especially in the age group of 16 to 29. The fatality rate per 10, 000 vehicles appears to be one of the highest in the world Artificial Neural Network (ANN) is a powerful technique that has demonstrated remarkable success in the analysis of historical data and in predicting future trends in many engineering fields. This novel technique was used to analyse the number of car accidents in the Sultanate of Oman during the period from 1976 to 1990. Input for the model was carefully selected through examining the strength of the correlation between the number of accidents and several related variables such as population growth, gross domestic product, number of vehicles on the road, etc. Results indicate that 95.5{\%} of the variation in the number of accidents could be explained by the model. Predictions for the years 1991-1994 showed high accuracy (92{\%}). To further validate the model, principal component analysis (PCA) regression technique was used to fit the same data, and predictions for 1991-1994 were obtained. Statistical analysis of the results showed that the ANN model prediction gave a lower mean absolute percentage error (MAPE) compared to the PCA. In addition, it also gave a lower MSE and a higher R2 value. These results demonstrate that ANN provided closer predictions to the actual results than PCA.",
author = "Al-Alawi, {S. M.} and Ali, {G. A.} and Bakheit, {C. S.}",
year = "1996",
month = "6",
language = "English",
volume = "5",
pages = "118--127",
journal = "Road and Transport Research",
issn = "1037-5783",
publisher = "ARRB Transport Research Ltd.",
number = "2",

}

TY - JOUR

T1 - A novel approach for traffic accident analysis and prediction using Artificial Neural Networks

AU - Al-Alawi, S. M.

AU - Ali, G. A.

AU - Bakheit, C. S.

PY - 1996/6

Y1 - 1996/6

N2 - Traffic accidents are among the major causes of death in the Sultanate of Oman, especially in the age group of 16 to 29. The fatality rate per 10, 000 vehicles appears to be one of the highest in the world Artificial Neural Network (ANN) is a powerful technique that has demonstrated remarkable success in the analysis of historical data and in predicting future trends in many engineering fields. This novel technique was used to analyse the number of car accidents in the Sultanate of Oman during the period from 1976 to 1990. Input for the model was carefully selected through examining the strength of the correlation between the number of accidents and several related variables such as population growth, gross domestic product, number of vehicles on the road, etc. Results indicate that 95.5% of the variation in the number of accidents could be explained by the model. Predictions for the years 1991-1994 showed high accuracy (92%). To further validate the model, principal component analysis (PCA) regression technique was used to fit the same data, and predictions for 1991-1994 were obtained. Statistical analysis of the results showed that the ANN model prediction gave a lower mean absolute percentage error (MAPE) compared to the PCA. In addition, it also gave a lower MSE and a higher R2 value. These results demonstrate that ANN provided closer predictions to the actual results than PCA.

AB - Traffic accidents are among the major causes of death in the Sultanate of Oman, especially in the age group of 16 to 29. The fatality rate per 10, 000 vehicles appears to be one of the highest in the world Artificial Neural Network (ANN) is a powerful technique that has demonstrated remarkable success in the analysis of historical data and in predicting future trends in many engineering fields. This novel technique was used to analyse the number of car accidents in the Sultanate of Oman during the period from 1976 to 1990. Input for the model was carefully selected through examining the strength of the correlation between the number of accidents and several related variables such as population growth, gross domestic product, number of vehicles on the road, etc. Results indicate that 95.5% of the variation in the number of accidents could be explained by the model. Predictions for the years 1991-1994 showed high accuracy (92%). To further validate the model, principal component analysis (PCA) regression technique was used to fit the same data, and predictions for 1991-1994 were obtained. Statistical analysis of the results showed that the ANN model prediction gave a lower mean absolute percentage error (MAPE) compared to the PCA. In addition, it also gave a lower MSE and a higher R2 value. These results demonstrate that ANN provided closer predictions to the actual results than PCA.

UR - http://www.scopus.com/inward/record.url?scp=0343370766&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=0343370766&partnerID=8YFLogxK

M3 - Article

VL - 5

SP - 118

EP - 127

JO - Road and Transport Research

JF - Road and Transport Research

SN - 1037-5783

IS - 2

ER -