Road Accident Prevention System using Machine Learning

Road accidents are a major public safety concern, often resulting in injuries, fatalities, and significant economic loss. Estimating the seriousness of an accident can aid emergency responders and authorities in taking quicker action and enhancing traffic safety. Machine learning offers powerful tools to analyze accident data and make accurate predictions based on various factorssuch as vehicle type, road conditions, driver behavior, and more. This project uses machine learning to predict how s ever a road accident severity using ensemble classification techniques. The dataset is first preprocessed by handling missing values and encoding categorical variables using Label Encoder. To address class imbalance, the Synthetic Minority Over-sampling Technique (SMOTE) is applied, ensuring equal representation of severity classes. The resampled data is then split into training and testing sets. AdaBoost with Random Forest combines the boosting power of AdaBoost with the strong prediction ability of Random Forest to improve classification accuracy. This approach helps in making better predictions even when the original data is imbalanced. Each model's performance is evaluated based on accuracy and the results are compared to identify the most effective model. This model achieved an accuracy of 91.19%, showing its effectiveness in handling imbalanced data and predicting accident severity. The web interface was coded using HTML and CSS with the Flask framework being utilized to connect the trained ML models to the webpage.