A ROAD ACCIDENT PREDICTION MODEL USING DATA MINING TECHNIQUES

Abstract

Daily accident rates are rising at an alarming pace, mostly as a result of the skyrocketing availability of motor cars. These days, there are a lot of traffic mishaps and fatalities, thus the transportation department needs a way to predict how many accidents will happen in a certain time period so they can make informed judgments. To better understand this situation and develop strategies to mitigate accidents, it would be beneficial to examine their frequency of occurrence. It becomes apparent after a while that there is a certain pattern to the accidents that happen in a single region, even though most incidents are characterized by ambiguity. Making educated guesses about the frequency of accidents in a given region and creating models to do so may be aided by this regularity. Our research in this study has focused on the environmental elements that contribute to road conditions, the correlations between these variables, and the frequency and severity of road accidents. We used the Apriori algorithm and Support Vector Machines, two data mining tools, to build a model that can anticipate accidents. This research used publicly accessible records (from 2014 to 2017) on vehicle accidents in Bangalore. This study's findings may be valuable for a variety of stakeholders, such as public works agencies, contractors, and other automotive businesses, who can utilize the estimations to improve road and vehicle design.