本研究旨在改良Huang(2019) 所提出之樣條函數節點篩選機制，並將其應用於Poisson 樣條迴歸中，以提升模型在不同節點配置下的估計準確性與配適效能。由於節點位置與數量的選擇對模型結果有顯著影響，因此本研究將節點視為需優化的變數選取議題。
研究中採用log likelihood 值作為模型參數選擇的依據，以挑選出最佳的參數組合，並進一步利用ISE 值（Integrated Squared Error）評估模型在不同節點設定下的近似效能。此外，我們亦進行多組模擬實驗，探討不同節點設置對最終估計結果的影響。模擬結果顯示，所提方法在估計準確度以及計算時間方面相較於結合Lasso 篩選節點之演算法，均有優良的表現。整體而言，本研究不僅補足現有方法於Poisson 樣條中的限制，也為進行樣條迴歸分析提供一套更具實用性與效能的節點選擇準則。

This study improves the spline knot selection mechanism proposed by Huang (2019) and applies it to generalized linear Poisson regression models. Recognizing that knot location and number significantly affect model performance, we frame knot selection as an optimization problem.
Our approach uses log-likelihood to identify optimal parameter combinations, and evaluates model performance under various knot settings using the Integrated Squared Error (ISE). Through simulation studies, we assess the impact of initial knot configurations and compare our method to Lasso-based selection algorithms. Results show that our method offers higher estimation accuracy and computational efficiency. This work addresses key limitations in existing spline-based Poisson models and provides a practical, effective strategy for knot selection.

De Boor, C. (1978). A Practical Guide to Splines. Springer, New York.
Gressani, O. and Eilers, P. H. (2024). Griddy-gibbs sampling for bayesian p-splines models with poisson data. arXiv preprint arXiv:2406.03336. https://arxiv.org/abs/2406.03336.
Huang, T.-M. (2019). A knot selection algorithm for regression splines. In Proceedings of the 62th ISI World Statistics Congress, Contributed Paper Session - Volume 2, pages 372–377.
Huang, T.-M. (2020). A knot selection algorithm for splines in logistic regression. In Proceedings of the 2020 3rd International Conference on Mathematics and Statistics (ICoMS), pages 29–33.
Kaishev, V. K., Dimitrova, D. S., Haberman, S., and Verrall, R. J. (2016). Geometrically designed, variable knot regression splines. Computational Statistics, 31(3):1079–1105.
Lindstrom, M. J. (1999). Penalized estimation of free-knot splines. Journal of Computational and Graphical Statistics, 8(2):333–352.
McCullagh, P. and Nelder, J. A. (1989). Generalized Linear Models. Chapman and Hall, London.
Schoenberg, I. (1946). Contributions to the problem of approximation of equidistant data by analytic functions: Part a. on the problem of smoothing or graduation. a first class of analytic approximation formulae. Quarterly of Applied Mathematics, 4(1):45–99.
Yuan, Y., Chen, N., and Zhou, S. (2013). Adaptive B-spline knot selection using multiresolution basis set. IIE Transactions, 45(12):1263–1277.
Zhou, S. and Shen, X. (2001). Spatially adaptive regression splines and accurate knot selection schemes. Journal of the American Statistical Association, 96:247–259.