Optimization Algorithms based Feature Selection Method for the Classification of Accidents

  • K.S. Ramakrishnan, Dr. C. Jothi Venkateswaran
Keywords: Machine Learning, Road Transportation, Feature Extraction, Feature Selection, Classification, Accident Dataset.

Abstract

Automobile engineers and researchers have attempted to design and develop safer vehicles, yet traffic accidents are inescapable. If we construct effective prediction models capable of automatically classifying the type of injury severity of diverse traffic accidents, we may be able to uncover patterns involved in dangerous wrecks. These behavioural and roadway accident patterns can help policymakers implement traffic safety measures. We believe that, in order to achieve the largest potential accident reduction impacts with limited financial resources, interventions must be founded on objective and scientific assessments of accident causes and injury severity. Governments are prioritizing traffic safety as one of their top objectives nowadays. Given the importance of the subject, determining the causes of traffic accidents has become the primary goal for reducing the harm caused by traffic accidents. In this paper, a hybrid feature selection method is developed by combining the PSO and Lion Optimization algorithms to obtain important road traffic features for accident categorization. ANN is used to categorize roads into accident zones and no accident zones during the classification stage.

Published
2021-08-07
How to Cite
Dr. C. Jothi Venkateswaran, K. R. (2021). Optimization Algorithms based Feature Selection Method for the Classification of Accidents. Design Engineering, 7067-7082. Retrieved from http://thedesignengineering.com/index.php/DE/article/view/3223
Section
Articles