Traffic Accidents Prediction Using Ensemble Machine Learning Approach

Lakshme Gowda, Monisha (2020) Traffic Accidents Prediction Using Ensemble Machine Learning Approach. Masters thesis, Dublin, National College of Ireland.

Preview	PDF (Master of Science) Download (1MB) | Preview
Preview	PDF (Configuration manual) Download (2MB) | Preview

Abstract

Research paper highlights the significance of various classification strategies in determining the occurrences of traffic accidents that happened throughout the data collection of learning accidents. In this paper, the main focus is to determine the classification using ensemble machine learning models for predicting the accidents when and where it occurs. Historical data for the study is retrieved from the UK transport department website which is open source for the researches. The utilized dataset comes under the huge spatial data for which the clustering is better classification technique. For large number of spatial datasets demands different knowledge regarding the clustering. DBSCAN clustering is better technique as it is capable of clustering the arbitrary shape data which is major requirement for spatial dataset. As the main problem of the research falls in the classification category, but classification requires positive and negative points for classification. The obtained clusters are positive accident spots, hence negative samples are generated by considering the date, time and cluster. The classification model like Random Forest, Ensemble Logistic Regression, AdaBoost Classifier, XGBoost Classifier, Ensemble (DT, SVM, Logistic) for prediction. The Random forest and XGBoost classification performance are pretty good in classifying the accidents based on the place and time when compared to other models. Evaluation metrics explains machine learning models fit better for classifying the accidents. Even though the proposed approach is promising, this study can be improved or extended in future work. Furthermore, in this research can be implemented in the real time traffic accident control concerns for avoiding the future incidents.
Key words: Traffic Accidents, DBSCAN Clustering, Negative Samples, Random Forest, Ensemble Logistic Regression, AdaBoost Classifier, XGBoost Classifier, Ensemble (DT, SVM, LogisticRegession)

Item Type:	Thesis (Masters)
Subjects:	Q Science > QA Mathematics > Electronic computers. Computer science T Technology > T Technology (General) > Information Technology > Electronic computers. Computer science Q Science > QA Mathematics > Computer software T Technology > T Technology (General) > Information Technology > Computer software
Divisions:	School of Computing > Master of Science in Data Analytics
Depositing User:	Dan English
Date Deposited:	22 Jan 2021 15:16
Last Modified:	22 Jan 2021 15:16
URI:	https://norma.ncirl.ie/id/eprint/4453

Actions (login required)

View Item