Research on the intrinsic mechanism of national logistics hub selection

To address the lack of data-driven support for understanding the underlying mechanisms of national logistics hub selection, this paper proposes a set of indicators influencing the evaluation of national logistics hubs and constructs a national logistics hub selection model based on Principal Component Analysis (PCA) and a kernel-free Twin Support Vector Machine (reg-LSDWPTSVM). This model determines whether a city can successfully be selected as a national logistics hub. First, a comprehensive evaluation index system was developed, providing a scientific foundation for studying the underlying mechanisms of logistics hub selection. PCA was then applied to reduce the dimensionality of relevant data, enabling feature extraction, and the reg-LSDWPTSVM model was used to evaluate the logistics hub selection process. The model achieved an accuracy rate of 95.33%,, a precision rate of 97.56%, and a recall rate of 86.96%,, thereby validating the rationality of the evaluation system with data-driven analysis. Finally, using production-oriented logistics hubs as an example, the constructed model was applied to explore key areas of development for relevant cities. The same method was used to generalize the socioeconomic development priorities of other cities, providing valuable insights for the development of national logistics hub cities.