Passenger-Centric Railroad Traffic Forecasting with RNN-Based Prediction

Predicting the traffic in the railroad is the secret to successful operation of the railroad and contented passengers. RNN-based real-time predictive model. The model training was in the form of real-time tracking data, weather reports, past train schedules, and maintenance records. The model was evaluated based on such metrics as MAE, RMSE, and R2. These promising values are the MAE of 2.3 minutes and RMSE of 3.1 minutes of the testing set, which indicates the model is correct in predicting the traffic situation. Moreover, the explanatory power of the model is enormous, with a value of R2 of 0.89. The flexibility is also seen in the fact that the model can always work in many different environments, including in bad weather or even when there is a need to carry out some repairs. According to the comparative analysis, there is evident improvement that leads to the better performance of the proposed work with an MAE of 2.3 minutes, RMSE of 3.1 minutes, and an R2 of 0.92. The results obtained demonstrate the quality of an RNN-based solution to improve the predictive railroad traffic forecasting in real time to obtain valuable information that can help to improve the operation of railways and to guarantee the satisfaction of passengers.