Predictive Subdata Selection for Gaussian Process Modeling

以作者查詢圖書館館藏

、以作者查詢臺灣博碩士

、以作者查詢全國書目

、勘誤回報

、線上人數：7

、訪客IP：3.129.20.133

姓名

徐永霖(Yung-Lin Hsu) 查詢紙本館藏

畢業系所

統計研究所

論文名稱

(Predictive Subdata Selection for Gaussian Process Modeling)

相關論文

★ Optimal Multi-platform Designs Based on Two Statistical Approaches	★ Subdata Selection : A- and I-optimalities
★ On the Construction of Multi-Stratum Factorial Designs	★ A Compression-Based Partitioning Estimate Classifier
★ On the Study of Feedforward Neural Networks: an Experimental Design Approach	★ Bayesian Optimization for Hyperparameter Tuning with Robust Parameter Design
★ Unreplicated Designs for Random Noise Exploration	★ Optimal Designs for Simple Directed/Weighted Network Structures
★ Study on the Prediction Capability of Two Aliasing Indices for Gaussian Random Fields	★ Optimal Designs on Undirected Network Structures for Network-Based Models
★ Data Reduction for Subsample in Gaussian Process	★ Gaussian Process Modeling with Weighted Additive Kernels

檔案

[Endnote RIS 格式]

[Bibtex 格式]

[相關文章]

[文章引用]

[完整記錄]

[館藏目錄]

至系統瀏覽論文 (2025-6-30以後開放)

摘要(中)

高斯過程被廣泛地運用在電腦實驗的建模上，但是隨著資訊爆炸的時代來臨，資料量成長的速度遠遠超過硬體設備的進步，使得高斯過程的預測能力在樣本數很大時急遽下降。常見的解決方法是透過改進模型或是實驗設計降低樣本數。本篇論文提出一個座標交換演算法，在給定樣本中迭代找出最具有預測力（predictive）的子樣本（subdata），且透過數值模擬來展示只需使用這些子樣本就可以達到準確的預測結果。

摘要(英)

Gaussian processes are widely used for emulating expensive computer simulators. However, their use is limited under large-scale data due to numerical problems. Common solutions to this difficulty are either to modify the model or to design suitable computer experiments. In this paper, we propose an exchange algorithm to search for predictive subdata given a full dataset. We demonstrate
our method on several examples to show the performance between our method and other competitive methods. Furthermore, we show that the predictive subdata selected by our method achieves high prediction accuracy even though the full dataset is not space-filling in the experimental region.

關鍵字(中)

★ 實驗設計
★ 電腦實驗
★ 座標交換演算法

關鍵字(英)

★ experimental design
★ computer experiments
★ exchange algorithm

論文目次

Abstract………………………………………………………………………………1
1. Introduction……………………………………………………………1
2. Preliminaries…………………………………………………………3
3. Methodology………………………………………………………………4
4. Simulation Studies……………………………………………8
5. Real Data Analysis…………………………………………13
6. Conclusion………………………………………………………………16
Appendix……………………………………………………………………………17
References………………………………………………………………………19

參考文獻

Ba, S., & Joseph, V. R. (2012). Composite gaussian process models for emulationg expensive functions. J. Appl. Stat., 6 , 1838–1860.
Fedorov, V. V. (1972). Theory of optimal experiments. Academic Press New York.
Gramacy, R., & Apley, D. (2015). Local gaussian process approximation for large computer experiments. J. Comput. Graph. Statist., 24 , 561–578.
Jones, B., & Goos, P. (2007). A candidate-set-free algorithm for generating doptimal split-plot designs. J. R. Stat. Soc. Ser. C. Appl. Stat., 56 , 347–364.
Marrel, A., Iooss, B., Dorpe, F. V., & Volkova, E. (2008). An efficient methodology for modeling complex computer codes with gaussian processes. Comput. Statist. Data Anal., 52 , 4731 – 4744.
Meyer, R. K., & Nachtsheim, C. J. (1995). The coordinate-exchange algorithm for constructing exact optimal experimental designs. Technometrics, 37 , 60–69.
Peng, C.-Y., & Wu, C. F. J. (2014). On the choice of nugget in kriging modeling for deterministic computer experiments. J. Comput. Graph. Statist., 23 , 151–168.
Plumlee, M., & Joseph, V. R. (2018). Orthogonal gaussian process models. Statist. Sinica, 28 , 601–619.
Sacks, J., Schiller, S. B., & Welch, W. J. (1989). Designs for computer experiments. Technometrics, 31 , 41–47.
Sambo, F., Borrotti, M., & Mylona, K. (2014). A coordinate-exchange twophase local search algorithm for the D- and I-optimal designs of split-plot experiments. Comput. Statist. Data Anal., 71 , 1193–1207.
Santner, T. J., Williams, B. J., & Notz, W. I. (2018). The Design and Analysis of Computer Experiments. New York: Springer.
Sch¨obi, R., Sudret, B., & Wiart, J. (2015). Polynomial-chaos-based kriging. Int. J. Uncertain. Quantif., 5 , 171–193.
Sun, D. X., Wu, C. F. J., & Chen, Y. (1997). Optimal blocking schemes for 2n and 2n−p designs. Technometrics, 39 , 298–307.
Sung, C.-L., Wang, W., Plumlee, M., & Haaland, B. (2019). Multiresolution functional anova for large-scale, many-input computer experiments. J. Amer. Statist. Assoc, 0 , 1–12.
Zhao, Y., Amemiya, Y., & Hung, Y. (2018). Efficient gaussian process modeling using experimental design-based subagging. Statist. Sinica, 28 , 1459–1479.

指導教授

張明中(Ming-Chung Chang)

審核日期

2021-1-22

推文