Prediction of road traffic accident quantity: multi factor regression analysis

To address the limitations of traditional prediction methods—including low recall rates, poor prediction accuracy, and long task completion times—this study proposes a road traffic accident frequency prediction model based on multifactor regression analysis. The multi factor regression analysis model extracts the influencing factors of road traffic accidents by calculating complex relationships, collects influencing factor data using road sensors, removes outliers, and performs recovery processing on the data to obtain high-quality data. Using ARIMA and LSTM models to predict the number of road traffic accidents based on the recovered data, in order to ensure the accuracy of the prediction results. The experimental results show that the average recall rate of the factors affecting road traffic accidents using the proposed method is 97.28%, the accuracy of predicting the number of road traffic accidents is between 94.67% and 97.64%, and the average completion time of the prediction task is 1.11s.