Accident Severity Detection Using Machine Learning A Review

Abstract

One of the greatest challenges in today's world is traffic accidents. It results in fatalities, accidents, and property damage. Making a model that can accurately predict traffic accidents is difficult. The objective of this project is to create a classification system for injuries based on a set of influential factors, including the environment, vehicle speed, driver behaviour, etc. Using data related to traffic accidents, several algorithms are utilized, including AdaBoost, Logistic Regression (LR), Naive Bayes (NB), and Random Forests (RF). Some of the best algorithms are the most effective, including Random Forest, Naive Bayes, and Ada Boost. Compared to LR, NB, and AdaBoost, the RF algorithm performed better, with 75.5% accuracy. To employ various machine learning classification algorithms for traffic accident prediction, the goal of this study is to uncover the underlying causes of road traffic accidents. then decide which prediction model is most likely to help decrease these highway accidents. This paper's goal is to review different authentication procedures offered by numerous scholars around the world.