Machine Learning algorithms are utilized for accident risk prediction
Given the challenges of public safety today, research and analysis of real-time traffic and accident data to predict the risk of accidents is ubiquitous. Accident risk prediction can significantly improve public safety by warning the public. In this paper, the probability of accident risk for a given case with selected conditions is predicted. Features like weather, traffic volume, road conditions, time of the day, description of previous accidents are utilized from the dataset. Machine learning algorithms like Logistic Regression, Support Vector Machines (SVM), Decision Tree, Neural Networks are used and their results are compared to provide the best prediction.
In this step, several models are implemented on the preprocessed dataset. The models implemented include the following
- Logistic Regression
- K Nearest Neighbors (KNN)
- MLP Classifier
- Decision Tree
- Random Forest Classifier
- Ensemble Modeling
- Accuracy Score
- Confusion matrix
- Cross validation score