Prediction of Bus Passenger Traffic using Gaussian Process Regression

Area of Science:

• Data Science
• Statistical Modeling
• Machine Learning

Background:

• Accurate passenger traffic analysis is crucial for efficient bus scheduling and traffic management.
• Existing models may lack the precision required for optimal resource allocation and passenger comfort.
• Bayesian analysis offers a robust framework for recursive data estimation.

Purpose of the Study:

• To design and implement an advanced passenger traffic prediction model.
• To leverage Gaussian Process Regression (GPR) for enhanced predictive accuracy.
• To compare the performance of the proposed GPR model against established methods.

Main Methods:

• The passenger traffic data was modeled as a Poisson process.
• A fully Bayesian Gaussian Process Regression (GPR) model was developed using PyMC3 with Theano.
• The prior for the GPR model was defined using a Gamma distributed function.

Main Results:

• The proposed Gaussian Process Regression (GPR) model demonstrated superior performance.
• The GPR method outperformed both the Student-t process model and Kernel Ridge Regression (KRR).
• The model provides more accurate predictions for passenger traffic.

Conclusions:

• Gaussian Process Regression (GPR) is a highly effective method for passenger traffic prediction.
• The developed GPR model offers significant improvements over traditional approaches.
• This approach enhances efficiency in bus scheduling and traffic management systems.