Classifying collisions in road accidents using XGBOOST, CATBOOST and SALP SWARM based optimization algorithms

Traffic accidents are the leading cause of death and injury in many developed nations. Anyone utilizing the road can meet an accident at any moment of time. The type of collision also plays a role in determining who is accountable for the accident. The biggest advantage of classifying collisions in...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Veröffentlicht in:Multimedia tools and applications Jg. 83; H. 13; S. 38387 - 38410
Hauptverfasser: Altaf, Insha, Kaul, Ajay
Format: Journal Article
Sprache:Englisch
Veröffentlicht: New York Springer US 01.04.2024
Springer Nature B.V
Schlagworte:
Algorithms
Classification
Collisions
Computer Communication Networks
Computer Science
Data Structures and Information Theory
Datasets
Feature selection
Multimedia Information Systems
Pedestrians
Roads & highways
Special Purpose and Application-Based Systems
Traffic accidents
Traffic accidents & safety
Type of collision
Improved salp swarm algorithm
Time-varying binary salp swarm algorithm
Enhanced salp swarm algorithm
Nature inspired algorithm
ISSN:1573-7721, 1380-7501, 1573-7721
Online-Zugang:Volltext
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
Beschreibung
Zusammenfassung:Traffic accidents are the leading cause of death and injury in many developed nations. Anyone utilizing the road can meet an accident at any moment of time. The type of collision also plays a role in determining who is accountable for the accident. The biggest advantage of classifying collisions in road accidents can pave a way for safer roads and reduced accident rates. A novel approach is proposed for classifying the type of collisions that might take place between vehicles and near by pedestrians, obstacles etc. on roads. A total of six hybrid classifiers are introduced in this article namely ` ` XGBoost c l a s s i f i e r u s i n g I S S A ′ ′ , ` ` XGBoost c l a s s i f i e r u s i n g E S S A ′ ′ , ` ` XGBoost c l a s s i f i e r u s i n g TVBSSA ′ ′ , ` ` CatBoost c l a s s i f i e r u s i n g I S S A ′ ′ , ` ` CatBoost c l a s s i f i e r u s i n g E S S A ′ ′ , and ` ` CatBoost c l a s s i f i e r u s i n g TVBSSA ′ ′ , The dataset considered in this article is the SWITRS dataset for classifying ` ` T y p e _ o f _ C o l l i s i o n ′ ′ . A total of 103000 accidents are considered when determining the ` ` T y p e _ o f _ C o l l i s i o n ′ ′ . It classifies the type of collisions using XGBoost algorithm, CatBoost Algorithm and three Nature Inspired Algorithms (NIA’s) have been used at the feature selection stage. The NIA’s considered for feature selection includes Improved Salp Swarm Algorithm (ISSA), Enhanced Salp Swarm Algorithm (ESSA), and Time-Varying Binary Salp Swarm Algorithm (TVBSSA). It is concluded that XGBoost c l a s s i f i e r u s i n g I S S A presents good stability with fewer hyper-parameters and the highest accuracy under different levels of training data volume. The value of Accuracy, Mean Square Error, and ROC-Auc in XGBoost using ISSA is 90.40, 0.1624 and 97.75, respectively. Moreover, the confusion matrix and evaluation metrics of XGBoost c l a s s i f i e r u s i n g I S S A performed better than the other two approaches. The findings of this study would be helpful in classifying the “type of collision”. These findings are highly significant in smart city projects to effectively establish timely proactive strategies and improve road traffic safety.
Bibliographie:ObjectType-Article-1
SourceType-Scholarly Journals-1
ObjectType-Feature-2
content type line 14
ISSN:1573-7721
1380-7501
1573-7721
DOI:10.1007/s11042-023-16969-4