From Basic to Advanced: Linear Regression Approaches for Analyzing Transportation Data

This paper employs both simple linear regression and multiple linear regression models to analyze transportation operation-related data. It primarily addresses three key issues: civil aviation business evaluation and passenger flow prediction, fuel consumption and payload prediction in aviation, and traffic accident impact forecasting. For civil aviation business evaluation and passenger flow prediction, the study applies multiple linear regression to assess the influence of factors such as national income on air passenger volume. While the constructed model demonstrates good fitting performance, limitations include a small sample size and multicollinearity among independent variables. Regarding traffic accident impact prediction, the research focuses on highway accidents by integrating VISSIM simulation, nonlinear regression, and neural networks to forecast spatiotemporal accident effects. Results indicate that the neural network achieves higher accuracy, whereas nonlinear regression offers the advantage of formula interpretability. Concerning fuel consumption and payload prediction, a simple linear regression model is established to quantify the relationship between aircraft payload and fuel usage. This provides dispatchers with actionable insights for fuel management and can be extended to different aircraft types and flight segments. Collectively, these studies offer significant methodological frameworks and decision-support tools for management across various transportation sectors.