Citation

P B, Goravanakolla S. Int. Res. J. Modern. Eng. Technol. Sci. 2023; 5(8): 1767-1772.

Copyright

(Copyright © 2023, IRJMETS)

DOI

10.56.726/IRJMETS44198

PMID

unavailable

Abstract

Single of the main causes of fatalities, disabling injuries, and hospitalisation in the nation continues to be traffic accidents. Because of this, predicting the likelihood of a traffic accident is crucial for preventing them and saving lives. The similar object has been addressing by a number of models, ranging from traditional models to contemporary approaches influenced by the increase of machine learning. In a stab to investigate and choose a practical strategy for travel disaster danger prediction, this paper examines many of these models. Given that drivers are in charge of the road, the study's goal is to estimate the likelihood of a traffic collision by looking at data that drivers would already be aware of, such as vehicle type, age, gender, time of day, weather, and so forth. The use of Optimum Identification Trees is an approach that, when combined with Random Forest and Logistic Regression, produces outcomes which make sense naturally to the user. In addition, geo-location data analysis utilising the K-means clustering technique can offer information about high-accident locations.

The road has gotten more difficult in the design and management sectors as the no of cars on the road has increased. Traffic accidents are a major source of worry on a worldwide scale because They significantly affect people's wellbeing, health, and safety. According to the World Health Organisation (WHO), 1.35 million people each year pass away in car accidents. As a result, they represent a substantial field of research for the use of cutting-edge methodologies and algorithms for analysis and prediction. While lots of street accidents are cause by external factors, some are cause by the driver. Unfavourable weather conditions, such as rain, clouds, and fog, for example, impair visibility and make driving on such roads difficult and often deadly. The present system prediction model evaluated only many probable causal factors.

Keywords: Road Accident, Traffic Accident, Machine Learning, K-Means, Geo-Location

Language: en