建議

經過上述的實例驗證證明利用本方法確實可以找到重要的自變數，建構精簡 但準確的預測模型。對於未來研究建議如下：

1. 支援向量機的變數影響性分析

本方法雖可找到支援向量機的重要變數，但無法判斷自變數對因變數是正 比、反比或曲線關係？自變數之間的交互作用對因變數是何關係？此一問題是否 也可用實驗計劃法解決一個值得研究的課題。

2. 支援向量機的參數設定之改進

本方法的確可以找到支援向量機的重要變數，但支援向量機的參數設定問題

仍是一項重要課題。本法在第一段找到重要變數後，在第二階段使用網格法找出

最佳懲罰係數、核心係數組合，但網格法並不是十分有效率的方法。是否可以從

第一階段的實驗數據中推估最佳參數的值是一個值得研究的方向。

參考文獻

1. Vapnik, V.N., The Nature of Statistical Learning Theory, Springer-Verlag, New York (1995).

2. Cortes, F., and Vapnik, V., “Support Vector Networks,” Machine Learning, Vol. 20, No. 3, pp.273-297 (1995).

3. Burges, C., “A tutorial on support vector machines for pattern recognition,” Data Mining and Knowledge Discovery, Vol. 2, No. 2, pp.121-167 (1998).

4. Cristianini, N., and Shawe-Taylor, J. An Introduction to Support Vector Machines, Cambridge University Press, Cambirdge University, 2000.

5. 柳回春、馬樹元，「支援向量機的研究現狀」，中國圖像圖形學報：A 輯，

第 7 卷，第 6 期，第 618-623 頁(2002)。

6. Lin, C.-J. “Formulations of support vector machines: a note from an optimization point of view,” Neural Computation, Vol.13, No.2, pp.307-317 (2001).

7. Freund, Y., and Schapire, R. E. “Large margin classification using the perceptron algorithm,” Machine Learning, Vol.37, No.3, pp.277-296 (2004).

8. Chung, K.-M., Ka, W.-C., Sun, C.-L., Wang, L.-L., and Lin, C.-J., “Radius margin bounds for support vector machines with the RBF kernel,” Neural Computation, Vol.15, pp.2643-2681 (2003).

9. Guyon, I., Weston, J., Barnhill, S., and Vapnik, V., “Gene selection for cancer classification using support vector machines,” Machine Learning, Vol.46, pp.389-422 (2002).

10. Drucker, H., Wu, D., and Vapink V., “Support vector machines for spam categorization,” IEEE Transactions on Neural Networks, Vol.10, No. 5, pp.1048-1054 (1999).

11. Hsu, C. W., Chang, C. C., & Lin, C. J., A practical guide to support vector classification, http://www.csie.ntu.edu.tw/~cjlin/papers/guide/guide.pdf (2003).

12. 黃建銘，「支撐向量機的自動模型選擇」，國立台灣科技大學資訊工程系，碩 士論文(2004)。

13. Chapelle, O., Vapnik, V., Bousquet, O, and Mukherjee, S. “Choosing multiple

parameters for support vector machines,” Machine Learning, Vol.46, pp.131-159

(2002).

14. Chen, Y.W. and Lin, C.J., “Combining SVMs with various feature selection strategies,” In I. Guyon, S. Gunn, M. Nikravesh, and L. Zadeh, editors, Feature extraction, foundations and Applications, Springer (2004).

15. Weston, J., Mukherjee, S., Chapelle, O., Pontil, M., Poggio, T., and Vapnik V.,

“Feature selection for SVMs,” Advances in Neural Information Processing Systems, Vol.12, pp.668-674 (2000).

16. Perkins, S., Lacker, K., and Theiler, J., “Grafting: Fast, incremental feature selection by gradient descent in function space,” Journal of Machine Learning Research, Vol.3, pp.1333-1356 (2003).

17. Guyon, I., and Elisseeff, A., “An introduction to variable and feature selection,”

Journal of Machine Learning Research, Vol. 3, pp.1157-1182 (2003).

18. Lemaire, V., and Clerot, F. “An input variable importance definition based on empirical data probability and its use in variable selection,” Proceedings of 2004 IEEE International Joint Conference on Neural Networks, Vol.2, pp.1375-1380 (2004).

19. Myers, R. H., and Montgomery, D. C., Response Surface Methodology, John Wiley & Sons, Inc., New York (1995).

20. Chang, C.C., and C.-J. Lin. LIBSVM: a library for support vector machines, 2001.

Software available at http://www.csie.ntu.edu.tw/~cjlin/libsvm.

21. Blackard, J.A., “Comparison of Neural Networks and Discriminant Analysis in Predicting Forest Cover Types,” Ph.D. Dissertation, Department of Forest Sciences, Colorado State University. Fort Collins, Colorado (1998).

22. Blake , C. L., and Merz , C. J., UCI repository of machine learning databases.

Technical report, University of California, Department of Information and Computer Science, Irvine, CA, 1998. Available at

http://kdd.ics.uci.edu/databases/covertype/covertype.html

23. 程韋綸，「倒傳遞網路的敏感性分析與灰箱倒傳遞神經網路」，中華大學，資 訊管理系，碩士論文(2007)。

24. Yeh, I-Cheng, “Modeling of strength of high performance concrete using artificial neural networks.” Cement and Concrete Research, Vol.28, No.12, pp.1797-1808 (1998).

25. ftp://ftp.ics.uci.edu/pub/machine-learning-databases/statlog/heart/http://www.ics.u ci.edu/~mlearn/MLSummary.html

26. Kuan-yu Chen., “Application of support vector regression in forecasting

international tourism demand,” Tourism Management Research, Vol.4, pp.81-97 (2004).

27. ftp://ftp.ics.uci.edu/pub/machine-learning-databases/thyroid-disease

附錄一：32 回合實驗設計表

表 1 七個變數實驗設計表

X1 X2 X3 X4 X5 X6 X7 1 -1 -1 -1 -1 -1 1 1 2 1 -1 -1 -1 -1 -1 -1 3 -1 1 -1 -1 -1 -1 -1 4 1 1 -1 -1 -1 1 1 5 -1 -1 1 -1 -1 -1 1 6 1 -1 1 -1 -1 1 -1 7 -1 1 1 -1 -1 1 -1 8 1 1 1 -1 -1 -1 1 9 -1 -1 -1 1 -1 -1 -1 10 1 -1 -1 1 -1 1 1 11 -1 1 -1 1 -1 1 1 12 1 1 -1 1 -1 -1 -1 13 -1 -1 1 1 -1 1 -1 14 1 -1 1 1 -1 -1 1 15 -1 1 1 1 -1 -1 1 16 1 1 1 1 -1 1 -1 17 -1 -1 -1 -1 1 1 -1 18 1 -1 -1 -1 1 -1 1 19 -1 1 -1 -1 1 -1 1 20 1 1 -1 -1 1 1 -1 21 -1 -1 1 -1 1 -1 -1 22 1 -1 1 -1 1 1 1 23 -1 1 1 -1 1 1 1 24 1 1 1 -1 1 -1 -1 25 -1 -1 -1 1 1 -1 1 26 1 -1 -1 1 1 1 -1 27 -1 1 -1 1 1 1 -1 28 1 1 -1 1 1 -1 1 29 -1 -1 1 1 1 1 1 30 1 -1 1 1 1 -1 -1 31 -1 1 1 1 1 -1 -1

32 1 1 1 1 1 1 1

表 2 八個變數實驗設計表

X1 X2 X3 X4 X5 X6 X7 X8 1 -1 -1 -1 -1 -1 -1 -1 1 2 1 -1 -1 -1 -1 1 1 1 3 -1 1 -1 -1 -1 1 1 -1 4 1 1 -1 -1 -1 -1 -1 -1 5 -1 -1 1 -1 -1 1 -1 -1 6 1 -1 1 -1 -1 -1 1 -1 7 -1 1 1 -1 -1 -1 1 1 8 1 1 1 -1 -1 1 -1 1 9 -1 -1 -1 1 -1 -1 1 -1 10 1 -1 -1 1 -1 1 -1 -1 11 -1 1 -1 1 -1 1 -1 1 12 1 1 -1 1 -1 -1 1 1 13 -1 -1 1 1 -1 1 1 1 14 1 -1 1 1 -1 -1 -1 1 15 -1 1 1 1 -1 -1 -1 -1

16 1 1 1 1 -1 1 1 -1

17 -1 -1 -1 -1 1 -1 -1 -1 18 1 -1 -1 -1 1 1 1 -1 19 -1 1 -1 -1 1 1 1 1 20 1 1 -1 -1 1 -1 -1 1 21 -1 -1 1 -1 1 1 -1 1 22 1 -1 1 -1 1 -1 1 1 23 -1 1 1 -1 1 -1 1 -1 24 1 1 1 -1 1 1 -1 -1 25 -1 -1 -1 1 1 -1 1 1 26 1 -1 -1 1 1 1 -1 1 27 -1 1 -1 1 1 1 -1 -1 28 1 1 -1 1 1 -1 1 -1 29 -1 -1 1 1 1 1 1 -1 30 1 -1 1 1 1 -1 -1 -1 31 -1 1 1 1 1 -1 -1 1

表 3 九~十個變數實驗設計表

X1 X2 X3 X4 X5 X6 X7 X8 X9 X10 1 -1 -1 -1 -1 -1 1 1 1 1 1 2 1 -1 -1 -1 -1 -1 -1 -1 -1 1 3 -1 1 -1 -1 -1 -1 -1 -1 1 -1 4 1 1 -1 -1 -1 1 1 1 -1 -1 5 -1 -1 1 -1 -1 -1 -1 1 -1 -1 6 1 -1 1 -1 -1 1 1 -1 1 -1 7 -1 1 1 -1 -1 1 1 -1 -1 1 8 1 1 1 -1 -1 -1 -1 1 1 1 9 -1 -1 -1 1 -1 -1 1 -1 -1 -1 10 1 -1 -1 1 -1 1 -1 1 1 -1 11 -1 1 -1 1 -1 1 -1 1 -1 1 12 1 1 -1 1 -1 -1 1 -1 1 1 13 -1 -1 1 1 -1 1 -1 -1 1 1 14 1 -1 1 1 -1 -1 1 1 -1 1 15 -1 1 1 1 -1 -1 1 1 1 -1 16 1 1 1 1 -1 1 -1 -1 -1 -1 17 -1 -1 -1 -1 1 1 -1 -1 -1 -1 18 1 -1 -1 -1 1 -1 1 1 1 -1 19 -1 1 -1 -1 1 -1 1 1 -1 1 20 1 1 -1 -1 1 1 -1 -1 1 1 21 -1 -1 1 -1 1 -1 1 -1 1 1 22 1 -1 1 -1 1 1 -1 1 -1 1 23 -1 1 1 -1 1 1 -1 1 1 -1 24 1 1 1 -1 1 -1 1 -1 -1 -1 25 -1 -1 -1 1 1 -1 -1 1 1 1 26 1 -1 -1 1 1 1 1 -1 -1 1 27 -1 1 -1 1 1 1 1 -1 1 -1 28 1 1 -1 1 1 -1 -1 1 -1 -1 29 -1 -1 1 1 1 1 1 1 -1 -1 30 1 -1 1 1 1 -1 -1 -1 1 -1 31 -1 1 1 1 1 -1 -1 -1 -1 1

32 1 1 1 1 1 1 1 1 1 1

表 4 十一~三十一個變數實驗設計表（0 即代表 -1）

X 1

X 2

X 3

X 4

X 5

X 6

X 7

X 8

X 9

X 1 0

X 1 1

X 1 2

X 1 3

X 1 4

X 1 5

X 1 6

X 1 7

X 1 8

X 1 9

X 2 0

X 2 1

X 2 2

X 2 3

X 2 4

X 2 5

X 2 6

X 2 7

X 2 8

X 2 9

X 3 0

X 3 1

1 0 0 0 0 0 1 1 1 1 1 1 1 1 1 1 0 0 0 0 0 0 0 0 0 0 1 1 1 1 1 0

2 1 0 0 0 0 0 0 0 0 1 1 1 1 1 1 1 1 1 1 1 1 0 0 0 0 0 0 0 0 1 1

3 0 1 0 0 0 0 1 1 1 0 0 0 1 1 1 1 1 1 0 0 0 1 1 1 0 0 0 0 1 0 1

4 1 1 0 0 0 1 0 0 0 0 0 0 1 1 1 0 0 0 1 1 1 1 1 1 0 1 1 1 0 0 0

5 0 0 1 0 0 1 0 1 1 0 1 1 0 0 1 1 0 0 1 1 0 1 1 0 1 0 0 1 0 0 1

6 1 0 1 0 0 0 1 0 0 0 1 1 0 0 1 0 1 1 0 0 1 1 1 0 1 1 1 0 1 0 0

7 0 1 1 0 0 0 0 1 1 1 0 0 0 0 1 0 1 1 1 1 0 0 0 1 1 1 1 0 0 1 0

8 1 1 1 0 0 1 1 0 0 1 0 0 0 0 1 1 0 0 0 0 1 0 0 1 1 0 0 1 1 1 1

9 0 0 0 1 0 1 1 0 1 1 0 1 0 1 0 0 1 0 1 0 1 1 0 1 1 0 1 0 0 0 1

10 1 0 0 1 0 0 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 1 1 0 1 1 0 0

11 0 1 0 1 0 0 1 0 1 0 1 0 0 1 0 1 0 1 1 0 1 0 1 0 1 1 0 1 0 1 0

12 1 1 0 1 0 1 0 1 0 0 1 0 0 1 0 0 1 0 0 1 0 0 1 0 1 0 1 0 1 1 1

13 0 0 1 1 0 1 0 0 1 0 0 1 1 0 0 1 1 0 0 1 1 0 1 1 0 1 0 0 1 1 0

14 1 0 1 1 0 0 1 1 0 0 0 1 1 0 0 0 0 1 1 0 0 0 1 1 0 0 1 1 0 1 1

15 0 1 1 1 0 0 0 0 1 1 1 0 1 0 0 0 0 1 0 1 1 1 0 0 0 0 1 1 1 0 1

16 1 1 1 1 0 1 1 1 0 1 1 0 1 0 0 1 1 0 1 0 0 1 0 0 0 1 0 0 0 0 0

17 0 0 0 0 1 1 1 1 0 1 1 0 1 0 0 0 0 1 0 1 1 0 1 1 1 1 0 0 0 0 1

18 1 0 0 0 1 0 0 0 1 1 1 0 1 0 0 1 1 0 1 0 0 0 1 1 1 0 1 1 1 0 0

19 0 1 0 0 1 0 1 1 0 0 0 1 1 0 0 1 1 0 0 1 1 1 0 0 1 0 1 1 0 1 0

20 1 1 0 0 1 1 0 0 1 0 0 1 1 0 0 0 0 1 1 0 0 1 0 0 1 1 0 0 1 1 1

21 0 0 1 0 1 1 0 1 0 0 1 0 0 1 0 1 0 1 1 0 1 1 0 1 0 0 1 0 1 1 0

22 1 0 1 0 1 0 1 0 1 0 1 0 0 1 0 0 1 0 0 1 0 1 0 1 0 1 0 1 0 1 1

23 0 1 1 0 1 0 0 1 0 1 0 1 0 1 0 0 1 0 1 0 1 0 1 0 0 1 0 1 1 0 1

24 1 1 1 0 1 1 1 0 1 1 0 1 0 1 0 1 0 1 0 1 0 0 1 0 0 0 1 0 0 0 0

25 0 0 0 1 1 1 1 0 0 1 0 0 0 0 1 0 1 1 1 1 0 1 1 0 0 0 0 1 1 1 0

26 1 0 0 1 1 0 0 1 1 1 0 0 0 0 1 1 0 0 0 0 1 1 1 0 0 1 1 0 0 1 1

27 0 1 0 1 1 0 1 0 0 0 1 1 0 0 1 1 0 0 1 1 0 0 0 1 0 1 1 0 1 0 1

28 1 1 0 1 1 1 0 1 1 0 1 1 0 0 1 0 1 1 0 0 1 0 0 1 0 0 0 1 0 0 0

29 0 0 1 1 1 1 0 0 0 0 0 0 1 1 1 1 1 1 0 0 0 0 0 0 1 1 1 1 0 0 1

30 1 0 1 1 1 0 1 1 1 0 0 0 1 1 1 0 0 0 1 1 1 0 0 0 1 0 0 0 1 0 0

31 0 1 1 1 1 0 0 0 0 1 1 1 1 1 1 0 0 0 0 0 0 1 1 1 1 0 0 0 0 1 0

32 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1

附錄二：各例題的 32 回合實驗設計表數據

表 1 交互與二次分類函數的 32 回合實驗設計表數據

C 0.001 10000 3 3 3 0.1 0.1 100 100 g 0.3 0.3 0.0001 1000 0.3 0.01 10 0.01 10

Y

1 57% 55% 57% 57% 57% 57% 57% 57% 51% 57%

2 57% 57% 57% 55% 57% 57% 57% 57% 57% 57%

3 57% 57% 57% 55% 57% 57% 57% 57% 57% 57%

4 57% 70% 57% 57% 65% 57% 57% 57% 62% 70%

5 57% 57% 57% 50% 57% 57% 57% 57% 57% 57%

6 57% 55% 57% 57% 57% 57% 57% 57% 56% 57%

7 57% 57% 57% 57% 57% 57% 57% 57% 51% 57%

8 57% 70% 57% 57% 64% 57% 57% 57% 67% 70%

9 57% 57% 57% 58% 57% 57% 57% 57% 57% 58%

10 57% 69% 57% 57% 69% 57% 57% 57% 70% 70%

11 57% 66% 57% 57% 64% 57% 57% 57% 64% 66%

12 57% 65% 57% 57% 68% 57% 57% 57% 63% 68%

13 57% 68% 57% 57% 62% 57% 57% 57% 67% 68%

14 57% 82% 57% 57% 78% 57% 57% 57% 78% 82%

15 57% 59% 57% 57% 57% 57% 57% 57% 51% 59%

16 57% 60% 57% 57% 57% 57% 57% 57% 54% 60%

17 57% 57% 57% 53% 57% 57% 57% 57% 55% 57%

18 57% 71% 57% 57% 69% 57% 58% 57% 66% 71%

19 57% 60% 57% 57% 57% 57% 57% 57% 58% 60%

20 57% 59% 57% 57% 57% 57% 57% 57% 61% 61%

21 57% 60% 57% 57% 57% 57% 57% 57% 62% 62%

22 57% 72% 57% 57% 66% 57% 57% 57% 75% 75%

23 57% 50% 57% 57% 57% 57% 57% 57% 61% 61%

24 57% 59% 57% 57% 57% 57% 57% 57% 63% 63%

25 57% 75% 57% 57% 66% 57% 57% 57% 79% 79%

26 57% 81% 57% 57% 67% 57% 57% 57% 72% 81%

17 57% 57% 57% 57% 57% 57% 57% 57% 56% 57%

28 57% 80% 57% 57% 68% 57% 59% 57% 76% 80%

29 57% 57% 57% 57% 57% 57% 57% 57% 56% 57%

30 57% 57% 57% 57% 57% 57% 57% 57% 66% 66%

31 57% 73% 57% 57% 68% 57% 58% 57% 72% 73%

32 57% 92% 57% 57% 84% 57% 57% 57% 63% 92%

表 2 交互與二次迴歸函數的 32 回合實驗設計表數據

C 0.001 10000 3 3 3 0.1 0.1 100 100 G 0.3 0.3 0.0001 1000 0.3 0.01 10 0.01 10

Y

1 0.0388 0.0424 0.0388 0.0387 0.0375 0.0387 0.0379 0.0383 0.0457 0.0375 2 0.0387 0.0371 0.0387 0.0738 0.0385 0.0387 0.0379 0.0386 0.0438 0.0371 3 0.0388 0.0391 0.0387 0.0621 0.0386 0.0388 0.0393 0.0388 0.0423 0.0386 4 0.0387 0.0226 0.0388 0.0387 0.0205 0.0387 0.0252 0.0263 0.0352 0.0205 5 0.0389 0.0394 0.0389 0.0536 0.0393 0.0390 0.0391 0.0389 0.0430 0.0389 6 0.0388 0.0525 0.0388 0.0387 0.0406 0.0387 0.0427 0.0398 0.0755 0.0387 7 0.0387 0.0435 0.0387 0.0387 0.0378 0.0386 0.0375 0.0379 0.0656 0.0375 8 0.0386 0.0231 0.0388 0.0387 0.0209 0.0387 0.0270 0.0262 0.0255 0.0209 9 0.0388 0.0389 0.0388 0.0632 0.0387 0.0387 0.0391 0.0385 0.0396 0.0385 10 0.0386 0.0276 0.0389 0.0387 0.0215 0.0389 0.0257 0.0265 0.0456 0.0215 11 0.0385 0.0242 0.0388 0.0387 0.0203 0.0387 0.0250 0.0239 0.0274 0.0203 12 0.0385 0.0410 0.0387 0.0387 0.0206 0.0386 0.0270 0.0237 0.0250 0.0206 13 0.0385 0.0223 0.0387 0.0387 0.0191 0.0386 0.0218 0.0236 0.0260 0.0191 14 0.0384 0.0079 0.0388 0.0387 0.0060 0.0387 0.0179 0.0119 0.0149 0.0060 15 0.0389 0.0646 0.0389 0.0387 0.0394 0.0388 0.0421 0.0393 0.0562 0.0387 16 0.0388 0.0416 0.0388 0.0387 0.0416 0.0388 0.0387 0.0399 0.0720 0.0387 17 0.0387 0.0314 0.0388 0.0526 0.0310 0.0388 0.0314 0.0345 0.0316 0.0310 18 0.0385 0.0155 0.0388 0.0387 0.0143 0.0389 0.0183 0.0224 0.0220 0.0143 19 0.0388 0.0328 0.0388 0.0386 0.0305 0.0388 0.0335 0.0347 0.0561 0.0305 20 0.0387 0.0453 0.0387 0.0387 0.0334 0.0386 0.0362 0.0353 0.0392 0.0334 21 0.0387 0.0325 0.0387 0.0387 0.0299 0.0386 0.0327 0.0346 0.0527 0.0299 22 0.0386 0.0208 0.0389 0.0387 0.0151 0.0387 0.0227 0.0228 0.0229 0.0151 23 0.0388 0.0493 0.0389 0.0387 0.0321 0.0389 0.0335 0.0360 0.0388 0.0321 24 0.0388 0.0350 0.0388 0.0387 0.0319 0.0389 0.0346 0.0352 0.0565 0.0319 25 0.0385 0.0156 0.0388 0.0387 0.0136 0.0388 0.0170 0.0205 0.0160 0.0136 26 0.0385 0.0243 0.0387 0.0387 0.0149 0.0386 0.0225 0.0209 0.0201 0.0149 17 0.0387 0.0435 0.0388 0.0387 0.0334 0.0387 0.0347 0.0352 0.0436 0.0334 28 0.0386 0.0155 0.0389 0.0387 0.0145 0.0390 0.0174 0.0226 0.0216 0.0145 29 0.0388 0.0512 0.0388 0.0387 0.0350 0.0389 0.0372 0.0370 0.0438 0.0350 30 0.0388 0.0385 0.0388 0.0387 0.0319 0.0388 0.0339 0.0364 0.0469 0.0319 31 0.0384 0.0163 0.0387 0.0387 0.0129 0.0385 0.0186 0.0200 0.0233 0.0129 32 0.0384 0.0037 0.0388 0.0387 0.0037 0.0387 0.0350 0.0091 0.0329 0.0037

表 3 線性與二次分類函數的 32 回合實驗設計表數據

C 0.001 10000 3 3 3 0.1 0.1 100 100 G 0.3 0.3 0.0001 1000 0.3 0.01 10 0.01 10

Y

1 60% 81% 60% 60% 78% 60% 61% 60% 70% 81%

2 60% 73% 60% 67% 75% 60% 74% 60% 73% 75%

3 60% 63% 60% 52% 60% 60% 59% 60% 58% 63%

4 60% 61% 60% 60% 60% 60% 60% 60% 56% 61%

5 60% 61% 60% 56% 60% 60% 61% 60% 63% 63%

6 60% 60% 60% 60% 60% 60% 60% 60% 53% 60%

7 60% 73% 60% 60% 74% 60% 61% 60% 64% 74%

8 60% 74% 60% 60% 78% 60% 60% 60% 70% 78%

9 60% 58% 60% 58% 57% 60% 63% 60% 56% 63%

10 60% 64% 60% 60% 60% 60% 60% 60% 56% 64%

11 60% 77% 60% 60% 80% 60% 61% 62% 70% 80%

12 60% 73% 60% 60% 78% 60% 60% 60% 67% 78%

13 60% 78% 60% 60% 80% 60% 63% 59% 58% 80%

14 60% 72% 60% 60% 79% 60% 60% 63% 69% 79%

15 60% 61% 60% 60% 64% 60% 60% 60% 46% 64%

16 60% 65% 60% 60% 60% 60% 60% 59% 58% 65%

17 60% 70% 60% 63% 72% 60% 70% 70% 70% 72%

18 60% 72% 60% 60% 74% 60% 70% 70% 73% 74%

19 60% 88% 60% 61% 79% 60% 77% 72% 84% 88%

20 60% 84% 60% 60% 85% 60% 61% 73% 83% 85%

21 60% 86% 60% 60% 79% 60% 74% 75% 85% 86%

22 60% 82% 60% 60% 83% 60% 60% 72% 83% 83%

23 60% 72% 60% 60% 76% 60% 61% 73% 62% 76%

24 60% 70% 60% 60% 72% 60% 63% 70% 63% 72%

25 60% 93% 60% 60% 86% 60% 84% 75% 88% 93%

26 60% 84% 60% 60% 79% 60% 61% 74% 76% 84%

17 60% 70% 60% 60% 74% 60% 61% 69% 64% 74%

28 60% 70% 60% 60% 73% 60% 69% 74% 58% 74%

29 60% 68% 60% 60% 76% 60% 60% 73% 61% 76%

30 60% 74% 60% 60% 75% 60% 71% 74% 67% 75%

31 60% 88% 60% 60% 84% 60% 75% 75% 80% 88%

32 60% 92% 60% 60% 83% 60% 60% 75% 73% 92%

表 4 線性與二次迴歸函數的 32 回合實驗設計表數據

C 0.001 10000 3 3 3 0.1 0.1 100 100

g 0.3 0.3 0.0001 1000 0.3 0.01 10 0.01 10

Y

1 14.6939 7.2135 14.6960 14.7031 9.4456 14.6861 12.8464 13.8804 13.9980 7.2135 2 14.6945 7.5529 14.6979 12.0442 9.8126 14.6920 9.1868 14.0701 8.1044 7.5529 3 14.6955 13.7586 14.6958 16.3245 13.6729 14.6859 13.6616 14.3737 13.9329 13.6616 4 14.6960 16.9757 14.6948 14.6991 14.4501 14.6832 14.5406 14.4542 25.3898 14.4501 5 14.6953 14.1945 14.6947 16.7289 13.9593 14.6834 14.2210 14.3592 15.3225 13.9593 6 14.6963 15.2152 14.6960 14.6991 14.2343 14.6874 14.3902 14.5015 29.3236 14.2343 7 14.6901 11.5094 14.6902 14.7004 10.3172 14.6682 13.2174 13.7802 17.6418 10.3172 8 14.6882 8.3796 14.6880 14.6991 9.3367 14.6610 13.5969 13.6951 10.9748 8.3796 9 14.6835 14.5329 14.6811 17.0992 14.0494 14.6388 14.2770 14.0171 16.0335 14.0171 10 14.6838 14.4945 14.6790 14.6988 13.4418 14.6337 14.2067 14.0770 30.8308 13.4418 11 14.6766 7.2480 14.6730 14.6961 8.9277 14.6136 12.8307 13.3409 15.4455 7.2480 12 14.6786 6.9220 14.6741 14.6991 9.1811 14.6170 13.6575 13.2936 11.5177 6.9220 13 14.6784 7.0964 14.6747 14.6999 9.1580 14.6176 12.7630 13.4174 12.2662 7.0964 14 14.6776 7.6912 14.6736 14.6991 9.0113 14.6133 13.5990 13.2701 11.4697 7.6912 15 14.6790 13.3242 14.6715 14.6989 12.8048 14.6083 14.2899 13.6137 18.8440 12.8048 16 14.6797 15.7524 14.6735 14.6996 13.9324 14.6148 14.2345 13.8010 27.7380 13.8010 17 14.6017 10.6640 14.5862 12.2920 10.5479 14.3337 11.0145 10.3078 11.6467 10.3078 18 14.6109 9.3766 14.5842 14.7094 9.4584 14.3289 12.2903 10.2503 18.4295 9.3766 19 14.6031 3.3336 14.5783 14.6049 4.8177 14.3095 10.6337 9.5677 8.2193 3.3336 20 14.6101 2.6978 14.5795 14.6991 4.4517 14.3138 12.6270 9.6977 4.1857 2.6978 21 14.6054 1.9342 14.5796 14.6968 3.7839 14.3141 10.4986 9.2764 4.2989 1.9342 22 14.6081 3.6174 14.5785 14.6991 4.3219 14.3105 12.5699 9.0927 5.8731 3.6174 23 14.6106 11.5328 14.5764 14.6991 8.6113 14.3045 13.3651 9.4901 13.7065 8.6113 24 14.6069 9.9782 14.5784 14.6771 9.4504 14.3110 12.7226 9.6506 20.1912 9.4504 25 14.5912 0.6216 14.5617 14.6990 2.1433 14.2559 9.8993 8.4875 1.6971 0.6216 26 14.5984 2.7314 14.5637 14.6991 4.2526 14.2634 12.3844 8.5342 4.3955 2.7314 17 14.6002 10.4159 14.5617 14.6986 8.1995 14.2577 13.2324 8.8644 14.0578 8.1995 28 14.5932 9.7832 14.5607 14.6984 8.8655 14.2523 12.2648 8.8999 19.1538 8.8655 29 14.5993 10.6649 14.5607 14.6995 8.5645 14.2534 13.2591 8.4272 12.7166 8.4272 30 14.5941 7.4156 14.5616 14.7001 7.4319 14.2564 12.0327 8.2827 13.6356 7.4156 31 14.5870 2.2314 14.5559 14.6967 3.1715 14.2374 10.3701 7.7588 5.2795 2.2314 32 14.6100 0.0313 14.5540 14.6991 2.1146 14.2355 14.5554 7.8867 10.7106 0.0313

表 5 森林地表覆蓋類型問題的 32 回合實驗設計表數據

C 0.001 10000 3 3 3 0.1 0.1 100 100 g 0.3 0.3 0.0001 1000 0.3 0.01 10 0.01 10

Y

1 16.1% 68.5% 16.1% 23.2% 62.4% 43.9% 57.2% 61.4% 62.9% 68.5%

2 16.1% 68.9% 16.1% 69.5% 63.9% 37.8% 66.5% 64.6% 71.4% 71.4%

3 16.1% 50.6% 16.1% 36.6% 49.8% 33.6% 48.4% 49.5% 45.5% 50.6%

4 16.1% 75.4% 16.1% 57% 69.3% 43.5% 70.3% 66.0% 75.6% 75.6%

5 16.1% 55.1% 16.1% 26.2% 52.9% 26.1% 50.1% 52.2% 50.9% 55.1%

6 16.1% 69.9% 16.1% 42.7% 70.0% 22.2% 67.8% 69.3% 65.9% 70.0%

7 16.1% 45.8% 16.1% 21.9% 43.9% 17.4% 41.9% 42.6% 39.6% 45.8%

8 16.1% 78.1% 16.1% 21.5% 72.6% 29.6% 64.7% 71.4% 75.2% 78.1%

9 16.1% 67.2% 16.1% 20.0% 60.5% 44.4% 55.3% 56.6% 67.5% 67.5%

10 16.1% 76.5% 16.1% 26.3% 68.4% 34.6% 67.9% 66.9% 72.8% 76.5%

11 16.1% 55.5% 16.1% 25.3% 52.9% 34.2% 52.2% 50.2% 51.8% 55.5%

12 16.1% 78.4% 16.1% 19.7% 73.0% 46.2% 67.5% 69.9% 75.2% 78.4%

13 16.1% 56.0% 16.1% 20.7% 54.7% 28.2% 48.7% 48.9% 53.2% 56.0%

14 16.1% 74.4% 16.1% 18.5% 71.7% 32.4% 62.1% 71.5% 66.3% 74.4%

15 16.1% 62.7% 16.1% 18.7% 58.1% 26.0% 46.9% 56.8% 59.9% 62.7%

16 16.1% 77.0% 16.1% 17.2% 76.9% 48.8% 29.3% 74.3% 75.2% 77.0%

17 16.1% 70.2% 16.1% 19.3% 63.5% 26.8% 50.9% 60.8% 67.1% 70.2%

18 16.1% 76.6% 16.1% 28.3% 70.2% 42.3% 69.6% 68.6% 74.2% 76.6%

19 16.1% 45.3% 16.1% 23.7% 44.3% 17.6% 39.3% 41.4% 38.2% 45.3%

20 16.1% 79.9% 16.1% 20.3% 72.6% 38.1% 69.0% 69.0% 74.4% 79.9%

21 16.1% 58.7% 16.1% 20.1% 55.2% 39.1% 49.4% 49.9% 53.4% 58.7%

22 16.1% 75.8% 16.1% 19.3% 71.9% 41.5% 63.1% 71.7% 68.0% 75.8%

23 16.1% 57.0% 16.1% 18.8% 53.6% 33.7% 44.9% 49.7% 53.6% 57.0%

24 16.1% 77.4% 16.1% 17.1% 75.4% 44.5% 37.4% 74.0% 76.1% 77.4%

25 16.1% 63.8% 16.1% 21.2% 54.1% 25.7% 49.2% 50.8% 62.8% 63.8%

26 16.1% 77.9% 16.1% 16.9% 74.0% 23.5% 56.2% 71.3% 74.0% 77.9%

17 16.1% 51.2% 16.1% 21.9% 49.7% 21.1% 48.2% 47.9% 45.8% 51.2%

28 16.1% 76.9% 16.1% 16.8% 74.9% 36.4% 39.6% 72.6% 74.9% 76.9%

29 16.1% 56.9% 16.1% 22.3% 53.8% 39.9% 49.3% 49.9% 55.3% 56.9%

30 16.1% 74.6% 16.1% 17.0% 73.8% 42.5% 36.2% 72.3% 65.7% 74.6%

31 16.1% 66.4% 16.1% 17.3% 60.4% 40.8% 45.5% 55.7% 65.4% 66.4%

32 16.1% 77.7% 16.2% 16.4% 77.3% 49.7% 19.3% 75.9% 72.7% 77.7%

表 6 混凝土強度個案的 32 回合實驗設計表數據

C 0.001 10000 3 3 3 0.1 0.1 100 100

g 0.3 0.3 0.0001 1000 0.3 0.01 10 0.01 10

Y

1 0.01061 0.01091 0.01053 0.01076 0.01024 0.01023 0.01081 0.00926 0.01073 0.00926 2 0.00909 0.01324 0.00803 0.01015 0.00585 0.00671 0.00699 0.00555 0.00738 0.00555 3 0.01117 0.03565 0.01143 0.00937 0.01225 0.01246 0.01185 0.01328 0.03374 0.00937 4 0.00946 0.00744 0.00867 0.01082 0.00780 0.00776 0.00867 0.00784 0.01482 0.00744 5 0.01111 0.01308 0.01114 0.01003 0.01339 0.01153 0.01265 0.01392 0.02321 0.01003 6 0.00904 0.02391 0.00812 0.00917 0.00751 0.00644 0.00717 0.00630 0.01124 0.00630 7 0.01085 0.01707 0.01065 0.00969 0.00945 0.01073 0.00960 0.00944 0.01890 0.00944 8 0.00937 0.00663 0.00798 0.01001 0.00337 0.00628 0.00694 0.00411 0.00715 0.00337 9 0.01032 0.01065 0.01026 0.00965 0.01090 0.00970 0.01004 0.01004 0.02076 0.00965 10 0.00887 0.01513 0.00796 0.00925 0.00740 0.00662 0.00882 0.00652 0.01132 0.00652 11 0.01016 0.03475 0.00959 0.01048 0.00833 0.00819 0.01041 0.00935 0.01068 0.00819 12 0.00917 0.01030 0.00763 0.01025 0.00522 0.00653 0.00870 0.00501 0.00880 0.00501 13 0.00995 0.03946 0.00932 0.01023 0.00941 0.00812 0.00951 0.00873 0.00955 0.00812 14 0.00829 0.00476 0.00709 0.00995 0.00292 0.00540 0.00584 0.00400 0.00583 0.00292 15 0.01041 0.04174 0.01008 0.01011 0.01149 0.00944 0.01074 0.01106 0.03618 0.00944 16 0.00952 0.01989 0.00772 0.00931 0.00482 0.00606 0.00931 0.00583 0.00965 0.00482 17 0.01156 0.01485 0.01143 0.01765 0.01467 0.01216 0.01519 0.01417 0.01641 0.01143 18 0.00992 0.03227 0.00904 0.00960 0.00958 0.00877 0.00986 0.01013 0.01073 0.00877 19 0.01098 0.06035 0.01088 0.01048 0.01159 0.01111 0.01279 0.01253 0.01327 0.01048 20 0.00939 0.00672 0.00819 0.01099 0.00586 0.00723 0.00733 0.00572 0.01047 0.00572 21 0.01101 0.02257 0.01090 0.01048 0.01748 0.01143 0.01417 0.01449 0.01846 0.01048 22 0.00925 0.02985 0.00782 0.01052 0.00448 0.00647 0.00765 0.00461 0.00822 0.00448 23 0.01097 0.26768 0.01108 0.00992 0.01324 0.01140 0.01193 0.01208 0.01351 0.00992 24 0.00987 0.03651 0.00867 0.00963 0.00653 0.00740 0.00946 0.00637 0.00990 0.00637 25 0.01038 0.02920 0.01002 0.01092 0.01059 0.00988 0.01268 0.01037 0.01406 0.00988 26 0.00913 0.01078 0.00778 0.01063 0.00515 0.00665 0.00890 0.00623 0.00908 0.00515 17 0.01082 0.13753 0.01078 0.01026 0.01433 0.01070 0.01228 0.01674 0.01248 0.01026 28 0.00962 0.08097 0.00837 0.00970 0.00732 0.00760 0.00984 0.00726 0.01009 0.00726 29 0.01078 0.03759 0.01095 0.00970 0.01075 0.01201 0.01029 0.01127 0.01044 0.00970 30 0.00919 0.04118 0.00805 0.00986 0.00543 0.00636 0.00897 0.00721 0.00945 0.00543 31 0.01026 0.02169 0.00972 0.01076 0.01009 0.00872 0.01079 0.01062 0.01073 0.00872 32 0.00960 0.00363 0.00736 0.01061 0.00337 0.00583 0.00932 0.00324 0.00955 0.00324

表 7 集集大地震引致山崩個案的 32 回合實驗設計表數據

C 0.001 10000 3 3 3 0.1 0.1 100 100 g 0.3 0.3 0.0001 1000 0.3 0.01 10 0.01 10

Y

1 49.6% 74.6% 68.1% 49.8% 79.1% 68.1% 52.1% 76.4% 75.9% 79.1%

2 49.6% 69.0% 64.0% 49.8% 72.4% 64.9% 63.4% 69.8% 69.4% 72.4%

3 49.6% 71.1% 62.0% 49.6% 72.1% 64.1% 60.9% 70.5% 69.4% 72.1%

4 49.6% 71.4% 49.9% 49.5% 72.5% 64.4% 61.1% 70.3% 65.0% 72.5%

5 49.6% 78.8% 70.6% 49.5% 76.9% 70.1% 65.9% 76.0% 76.5% 78.8%

6 49.6% 69.6% 65.6% 49.5% 67.5% 64.9% 66.4% 68.3% 63.5% 69.6%

7 49.6% 73.4% 62.0% 55.5% 70.3% 59.5% 67.1% 69.1% 72.1% 73.4%

8 49.6% 69.1% 49.6% 49.4% 70.8% 59.4% 57.1% 67.5% 72.1% 72.1%

9 49.6% 77.5% 65.0% 50.3% 77.5% 66.1% 77.9% 75.8% 77.5% 77.9%

10 49.6% 75.9% 60.9% 50.3% 74.9% 64.9% 67.8% 72.9% 71.5% 75.9%

11 49.6% 72.1% 65.8% 49.9% 70.8% 67.1% 67.3% 69.4% 65.5% 72.1%

12 49.6% 74.6% 58.9% 49.6% 77.1% 67.6% 60.8% 76.1% 73.5% 77.1%

13 49.6% 76.3% 64.3% 51.0% 77.9% 64.5% 70.1% 75.9% 74.9% 77.9%

14 49.6% 74.8% 63.1% 49.8% 75.4% 63.4% 61.9% 72.1% 69.3% 75.4%

15 49.6% 71.0% 65.0% 49.8% 72.4% 65.5% 70.8% 69.8% 68.1% 72.4%

16 49.6% 73.1% 66.1% 49.6% 78.9% 66.5% 49.6% 76.8% 72.6% 78.9%

17 49.6% 77.1% 68.3% 50.8% 78.9% 68.6% 76.1% 75.5% 74.3% 78.9%

18 49.6% 75.4% 68.0% 50.9% 73.6% 68.9% 73.1% 71.3% 70.5% 75.4%

19 49.6% 73.1% 68.5% 49.9% 75.3% 69.9% 67.1% 76.8% 70.4% 76.8%

20 49.6% 76.0% 69.9% 49.8% 80.6% 70.6% 68.3% 76.1% 73.9% 80.6%

21 49.6% 78.9% 63.9% 50.6% 79.0% 68.8% 69.8% 76.3% 75.0% 79.0%

22 49.6% 71.3% 67.0% 49.5% 72.4% 69.1% 68.1% 71.9% 69.6% 72.4%

23 49.6% 75.6% 68.9% 49.5% 76.9% 69.6% 60.1% 75.4% 73.3% 76.9%

24 49.6% 73.8% 70.0% 49.5% 79.5% 70.4% 49.8% 77.3% 74.4% 79.5%

25 49.6% 77.5% 51.0% 58.8% 75.9% 63.6% 72.9% 74.8% 75.5% 77.5%

26 49.6% 79.5% 64.1% 57.5% 75.6% 64.8% 70.8% 72.9% 76.1% 79.5%

17 49.6% 72.9% 70.6% 49.5% 74.0% 71.5% 65.8% 74.9% 71.4% 74.9%

28 49.6% 76.8% 70.4% 49.5% 80.3% 72.4% 56.3% 78.8% 75.4% 80.3%

29 49.6% 77.5% 56.0% 50.0% 77.6% 67.3% 62.9% 75.3% 76.9% 77.6%

30 49.6% 73.3% 64.1% 49.6% 76.6% 65.1% 49.6% 74.1% 75.0% 76.6%

31 49.6% 72.0% 70.5% 49.6% 75.3% 72.4% 49.8% 74.4% 70.1% 75.3%

32 49.6% 77.5% 70.0% 49.6% 79.6% 71.8% 49.6% 77.9% 68.4% 79.6%

表 8 休旅車潛在客戶開發個案的 32 回合實驗設計表數據

C 0.001 10000 3 3 3 0.1 0.1 100 100 g 0.3 0.3 0.0001 1000 0.3 0.01 10 0.01 10

Y

1 7.0121 6.8020 6.4766 7.0046 6.8176 6.2241 7.0208 8.9297 7.0008 6.2241 2 7.0122 7.0068 6.7019 7.0046 6.9944 6.5117 7.0208 7.8880 7.0008 6.5117 3 7.0121 6.5781 6.2957 7.0046 6.6268 5.8417 7.0208 5.7885 7.0008 5.7885 4 7.0122 6.8756 6.1380 7.0046 6.8940 5.7876 7.0208 6.0065 7.0008 5.7876 5 7.0122 6.8747 6.3725 7.0046 6.8904 6.0183 7.0208 6.2927 7.0008 6.0183 6 7.0122 6.8799 6.1456 7.0046 6.8979 5.8033 7.0208 6.5721 7.0008 5.8033 7 7.0122 6.8780 6.0633 7.0046 6.8942 5.6239 7.0208 7.2387 7.0008 5.6239 8 7.0121 6.7008 5.7001 7.0046 6.7276 5.1933 7.0208 6.2522 7.0008 5.1933 9 7.0122 6.8822 6.2620 7.0046 6.8970 5.9577 7.0208 8.8115 7.0008 5.9577 10 7.0122 6.8728 5.8867 7.0046 6.8924 5.4252 7.0208 7.1890 7.0008 5.4252 11 7.0122 6.9334 6.1876 7.0046 6.9460 5.8238 7.0208 6.8422 7.0008 5.8238 12 7.0121 6.7807 6.1524 7.0046 6.8060 5.7076 7.0208 5.4964 7.0008 5.4964 13 7.0122 6.8844 6.1622 7.0046 6.9039 5.7508 7.0208 6.3257 7.0008 5.7508 14 7.0121 6.7734 6.1997 7.0046 6.7954 5.7422 7.0208 6.0170 7.0008 5.7422 15 7.0122 6.8251 6.4083 7.0046 6.8504 6.1497 7.0208 8.4873 7.0008 6.1497 16 7.0122 6.9529 6.5102 7.0046 6.9660 6.3360 7.0208 10.5147 7.0008 6.3360 17 7.0121 6.7877 5.8508 7.0046 6.8097 5.3977 7.0208 6.8732 7.0008 5.3977 18 7.0121 6.8199 5.6275 7.0046 6.8440 5.0896 7.0208 5.3398 7.0008 5.0896 19 7.0122 6.8583 5.8398 7.0046 6.8792 5.3290 7.0208 4.6787 7.0008 4.6787 20 7.0122 6.8232 5.7364 7.0046 6.8470 5.1703 7.0208 5.3004 7.0008 5.1703 21 7.0122 6.8893 5.8399 7.0046 6.9082 5.3400 7.0208 5.1517 7.0008 5.1517 22 7.0121 6.8069 5.7462 7.0046 6.8313 5.1038 7.0208 4.6143 7.0008 4.6143 23 7.0121 6.8151 5.9921 7.0046 6.8375 5.4763 7.0208 5.9069 7.0008 5.4763 24 7.0122 6.8929 6.2236 7.0046 6.9037 5.8539 7.0208 7.8196 7.0008 5.8539 25 7.0122 6.9555 6.0565 7.0046 6.9619 5.6986 7.0208 6.5987 7.0008 5.6986 26 7.0121 6.7763 6.1325 7.0046 6.8035 5.7108 7.0208 6.1928 7.0008 5.7108 17 7.0122 6.8525 5.8899 7.0046 6.8715 5.2949 7.0208 4.7590 7.0008 4.7590 28 7.0121 6.8012 5.8496 7.0046 6.8340 5.3189 7.0208 4.5864 7.0008 4.5864 29 7.0121 6.7468 5.9267 7.0046 6.7820 5.3662 7.0208 5.3729 7.0008 5.3662 30 7.0122 6.8724 5.9042 7.0046 6.8884 5.3836 7.0208 5.0096 7.0008 5.0096 31 7.0121 6.7687 5.7102 7.0046 6.7914 5.1916 7.0208 4.8000 7.0008 4.8000 32 7.0122 7.0005 5.2040 7.0046 7.0044 5.4944 7.0208 3.8538 7.0008 3.8538

表 9 心臟病診斷個案的 32 回合實驗設計表數據

C 0.001 10000 3 3 3 0.1 0.1 100 100 g 0.3 0.3 0.0001 1000 0.3 0.01 10 0.01 10

最大

機率 診斷成本 總成本

1 55.6% 74.4% 55.6% 54.8% 78.1% 60.4% 55.6% 81.5% 71.5% 81.5% 55200 240385 2 55.6% 71.9% 55.6% 68.5% 79.3% 58.9% 59.3% 78.1% 72.2% 79.3% 45010 252417 3 55.6% 73.3% 55.6% 69.3% 75.6% 55.6% 75.2% 76.7% 75.2% 76.7% 39210 272543 4 55.6% 78.1% 55.6% 72.2% 75.6% 55.6% 72.6% 75.6% 77.0% 78.1% 31020 249539 5 55.6% 75.2% 55.6% 61.1% 83.0% 55.6% 59.6% 81.5% 72.6% 83.0% 11600 181970 6 55.6% 68.9% 55.6% 67.0% 79.6% 55.6% 64.1% 81.5% 71.5% 81.5% 19410 204595 7 55.6% 75.6% 55.6% 58.9% 78.1% 55.6% 55.6% 77.8% 68.9% 78.1% 5610 224129 8 55.6% 69.3% 55.6% 58.1% 75.6% 55.6% 55.6% 77.8% 65.6% 77.8% 15420 237642 9 55.6% 70.0% 55.6% 55.2% 77.4% 56.7% 55.6% 78.1% 70.7% 78.1% 20200 238719 10 55.6% 70.7% 55.6% 55.6% 73.0% 56.7% 59.6% 74.8% 66.7% 74.8% 10210 262062 11 55.6% 71.5% 55.6% 66.7% 73.0% 55.6% 66.3% 67.4% 70.4% 73.0% 14210 284580 12 55.6% 66.7% 55.6% 54.8% 73.0% 55.6% 70.4% 71.9% 57.8% 73.0% 6220 276590 13 55.6% 78.9% 55.6% 66.7% 84.1% 55.6% 66.3% 81.9% 76.3% 84.1% 36600 195859 14 55.6% 72.6% 55.6% 56.3% 81.5% 55.6% 55.6% 84.1% 70.7% 84.1% 44610 203869 15 55.6% 73.0% 55.6% 55.9% 77.0% 57.8% 55.6% 78.5% 69.3% 78.5% 40610 255425 16 55.6% 74.4% 55.6% 55.6% 75.2% 56.7% 55.6% 80.4% 67.8% 80.4% 50620 246916 17 55.6% 74.1% 55.6% 55.9% 76.3% 55.9% 55.6% 77.0% 67.0% 77.0% 51100 280730 18 55.6% 71.5% 55.6% 55.2% 74.4% 56.7% 55.6% 77.0% 67.8% 77.0% 41110 270740 19 55.6% 74.8% 55.6% 55.6% 79.6% 55.6% 55.6% 78.5% 73.7% 79.6% 45110 248814 20 55.6% 73.3% 55.6% 57.8% 77.4% 55.6% 55.6% 80.4% 73.3% 80.4% 37120 233416 21 55.6% 72.2% 55.6% 60.0% 70.4% 55.6% 57.4% 75.2% 68.1% 75.2% 7500 255648 22 55.6% 67.8% 55.6% 58.9% 73.0% 55.6% 56.3% 74.1% 63.7% 74.1% 15510 274769 23 55.6% 74.4% 55.6% 55.9% 81.1% 56.7% 55.6% 82.2% 73.7% 82.2% 11510 189288 24 55.6% 71.5% 55.6% 55.2% 78.5% 57.0% 55.6% 81.1% 61.9% 81.1% 21520 210409 25 55.6% 63.3% 55.6% 55.2% 67.4% 55.6% 55.6% 68.5% 62.6% 68.5% 16100 330915 26 55.6% 71.9% 55.6% 55.6% 74.4% 55.6% 55.6% 75.2% 65.9% 75.2% 6310 254458 17 55.6% 70.7% 55.6% 55.9% 72.6% 55.6% 55.9% 74.8% 68.5% 74.8% 20110 271962 28 55.6% 74.1% 55.6% 55.6% 79.3% 55.6% 55.6% 81.1% 73.7% 81.1% 12320 201209 29 55.6% 72.6% 55.6% 60.0% 76.3% 55.6% 69.3% 74.4% 67.4% 76.3% 32500 269537 30 55.6% 63.0% 55.6% 55.6% 77.0% 55.6% 55.6% 78.9% 69.3% 78.9% 40710 251821 31 55.6% 76.3% 55.6% 55.6% 82.2% 61.5% 55.6% 84.1% 67.0% 84.1% 46510 205769 32 55.6% 83.3% 55.6% 55.6% 82.2% 62.6% 55.6% 84.1% 58.5% 84.1% 56720 215979

表 10 甲狀腺機能診斷個案的 32 回合實驗設計表數據

C 0.001 10000 3 3 3 0.1 0.1 100 100 g 0.3 0.3 0.0001 1000 0.3 0.01 10 0.01 10

最大 機率

診斷

成本 總成本

1 92.58% 92.56% 92.58% 92.56% 92.58% 92.58% 92.58% 92.58% 92.56% 92.58% 10 233 2 92.58% 95.04% 92.58% 93.96% 93.46% 92.58% 93.06% 93.49% 95.24% 95.24% 70 213 3 92.58% 94.36% 92.58% 95.68% 93.13% 92.58% 93.01% 93.22% 94.35% 95.68% 42 172 4 92.58% 93.63% 92.58% 93.49% 92.58% 92.58% 92.58% 92.58% 93.65% 93.65% 34 224 5 92.58% 93.57% 92.58% 93.54% 92.58% 92.58% 92.58% 92.58% 93.58% 93.58% 31 223 6 92.58% 96.81% 92.58% 96.22% 93.36% 92.58% 93.15% 93.38% 96.89% 96.89% 45 139 7 92.58% 97.68% 92.58% 97.85% 93.29% 92.58% 92.94% 93.36% 97.29% 97.85% 63 128 8 92.58% 93.64% 92.58% 93.69% 92.58% 92.58% 92.58% 92.58% 93.65% 93.69% 13 202 9 92.58% 94.81% 92.58% 96.22% 93.38% 92.58% 93.25% 93.46% 95.13% 96.22% 46 160 10 92.58% 92.53% 92.58% 92.65% 92.58% 92.58% 92.58% 92.58% 92.51% 92.65% 30 250 11 92.58% 93.82% 92.58% 93.81% 92.58% 92.58% 92.58% 92.58% 93.85% 93.85% 35 220 12 92.58% 94.39% 92.58% 93.81% 93.13% 92.58% 92.64% 93.22% 94.24% 94.39% 41 210 13 92.58% 96.90% 92.58% 97.82% 93.38% 92.58% 93.28% 93.36% 96.75% 97.82% 46 112 14 92.58% 93.54% 92.58% 93.39% 92.58% 92.58% 92.58% 92.58% 93.53% 93.54% 30 224 15 92.58% 93.74% 92.58% 93.64% 92.58% 92.58% 92.58% 92.58% 93.76% 93.76% 35 222 16 92.58% 97.14% 92.58% 94.69% 93.29% 92.58% 92.58% 93.33% 96.35% 97.14% 45 131 17 92.58% 93.68% 92.58% 93.69% 92.58% 92.58% 92.58% 92.58% 93.72% 93.72% 35 223 18 92.58% 94.96% 92.58% 94.71% 93.33% 92.58% 92.92% 93.36% 95.13% 95.13% 41 188 19 92.58% 94.74% 92.58% 96.35% 93.35% 92.58% 93.17% 93.36% 95.04% 96.35% 46 156 20 92.58% 93.58% 92.58% 93.42% 92.58% 92.58% 92.58% 92.58% 93.54% 93.58% 30 222 21 92.58% 93.76% 92.58% 93.75% 92.58% 92.58% 92.58% 92.58% 93.76% 93.76% 35 222 22 92.58% 96.14% 92.58% 95.31% 93.11% 92.58% 92.76% 93.19% 96.17% 96.17% 41 156 23 92.58% 97.58% 92.58% 97.58% 93.36% 92.58% 92.99% 93.46% 97.07% 97.58% 46 119 24 92.58% 92.56% 92.58% 92.40% 92.58% 92.58% 92.58% 92.58% 92.50% 92.58% 34 256 25 92.58% 94.74% 92.58% 96.28% 93.36% 92.58% 93.25% 93.36% 95.19% 96.28% 63 175 26 92.58% 93.64% 92.58% 93.53% 92.58% 92.58% 92.58% 92.58% 93.61% 93.64% 13 204 17 92.58% 93.67% 92.58% 93.58% 92.58% 92.58% 92.58% 92.58% 93.68% 93.68% 31 221 28 92.58% 95.17% 92.58% 94.33% 93.33% 92.58% 92.74% 93.38% 94.81% 95.17% 49 194 29 92.58% 96.19% 92.58% 97.33% 93.13% 92.58% 93.11% 93.21% 95.92% 97.33% 42 122 30 92.58% 93.51% 92.58% 93.33% 92.58% 92.58% 92.58% 92.58% 93.39% 93.51% 38 233 31 92.58% 92.54% 92.58% 92.54% 92.58% 92.58% 92.58% 92.58% 92.54% 92.58% 10 233 32 92.58% 97.44% 92.58% 94.04% 93.42% 92.58% 92.60% 93.50% 96.35% 97.44% 78 155

附錄三：各例題的網格法結果

表 11 例題一之網格法結果

核心係數 g

懲罰係數 C

0.0001 0.001 0.01 0.1 1 10 100 1000

0.01 57% 57% 57% 57% 57% 57% 57% 57%

0.1 57% 57% 57% 57% 60% 70% 57% 57%

1 57% 57% 57% 57% 91% 91% 57% 57%

10 57% 57% 57% 73% 95% 91% 63% 57%

100 57% 57% 57% 92% 95% 89% 63% 57%

1000 57% 57% 74% 94% 93% 89% 63% 57%

10000 57% 57% 91% 97% 94% 89% 63% 57%

100000 57% 73% 96% 97% 94% 89% 63% 57%

0. 0001 0. 001 0.01 0.1 1 10 100 1000

0.01 0.1 1 10 100 1000 10000 100000

核心係數g

懲罰係 數 C

95%-100%

90%-95%

85%-90%

80%-85%

75%-80%

70%-75%

65%-70%

60%-65%

55%-60%

50%-55%

圖 1 例題一之網格法結果 2D 圖

表 12 例題二之網格法結果

核心係數 g

懲罰係數C 0.0001 0.001 0.01 0.1 1 10 100 1000

0.01 0.03880 0.03880 0.03879 0.03811 0.01965 0.02291 0.03843 0.03879 0.1 0.03880 0.03880 0.03877 0.03268 0.00378 0.00808 0.03580 0.03896 1 0.03880 0.03882 0.03788 0.00907 0.00219 0.00618 0.03295 0.03868 10 0.03883 0.03852 0.03202 0.00230 0.00219 0.00618 0.03295 0.03868 100 0.03859 0.03772 0.00812 0.00191 0.00219 0.00618 0.03295 0.03868 1000 0.03835 0.03203 0.00229 0.00191 0.00219 0.00618 0.03295 0.03868 10000 0.03760 0.00798 0.00184 0.00191 0.00219 0.00618 0.03295 0.03868 100000 0.03176 0.00216 0.00184 0.00191 0.00219 0.00618 0.03295 0.03868

0.0 001 0. 001 0.01 0.1 1 10 100 1000

0.01 0.1 1 10 100 1000 10000 100000

核心係數g

懲罰係 數 C

0.035 -0.040 0.030 -0.035 0.025 -0.030 0.020 -0.025 0.015 -0.020 0.010 -0.015 0.005 -0.010 0.000 -0.005

圖 2 例題二之網格法結果 2D 圖

表 13 例題三之網格法結果

核心係數 g

懲罰係數C 0.0001 0.001 0.01 0.1 1 10 100 1000

0.01

60.0% 60.0% 60.0% 60.0% 60.0% 60.0% 60.0% 60.0%

0.1

60.0% 60.0% 60.0% 73.0% 79.0% 84.0% 60.0% 60.0%

1

60.0% 60.0% 73.0% 75.0% 92.0% 92.0% 61.0% 60.0%

10

60.0% 73.0% 74.0% 81.0% 95.0% 90.0% 65.0% 60.0%

100

73.0% 73.0% 75.0% 93.0% 91.0% 88.0% 65.0% 60.0%

1000

73.0% 73.0% 82.0% 93.0% 89.0% 88.0% 65.0% 60.0%

10000

72.0% 75.0% 93.0% 93.0% 89.0% 88.0% 65.0% 60.0%

100000

72.0% 82.0% 94.0% 93.0% 89.0% 88.0% 65.0% 60.0%

0.0 001 0. 001 0.01 0.1 1 10 100 1000

0.01 0.1 1 10 100 1000 10000 100000

核心係數g

懲罰係 數 C

90.0%-95.0%

85.0%-90.0%

80.0%-85.0%

75.0%-80.0%

70.0%-75.0%

65.0%-70.0%

60.0%-65.0%

55.0%-60.0%

50.0%-55.0%

圖 3 例題三之網格法結果 2D 圖

表 14 例題四之網格法結果

核心係數 g

懲罰係數C 0.0001 0.001 0.01 0.1 1 10 100 1000 0.01

14.70 14.70 14.65 14.29 12.85 14.11 14.70 14.70

0.1

14.70 14.65 14.26 11.76 7.25 9.90 14.65 14.70

1

14.65 14.25 11.65 8.49 1.30 2.45 14.20 14.70

10

14.25 11.64 8.95 4.77 0.72 1.46 12.73 14.71

100

11.64 9.00 8.49 0.69 0.69 1.70 12.73 14.71

1000

9.00 9.03 4.50 0.55 0.75 1.76 12.73 14.71

10000

9.09 8.50 0.67 0.53 1.01 1.76 12.73 14.71

100000

9.05 4.51 0.54 0.57 1.27 1.76 12.73 14.71

0.0 001 0. 001 0.01 0.1 1 10 100 1000

0.01 0.1 1 10 100 1000 10000 100000

核心係數g

懲罰係 數 C

14.0 -16.0 12.0 -14.0 10.0 -12.0 8.0 -10.0 6.0 -8.0 4.0 -6.0 2.0 -4.0 0.0 -2.0

圖 4 例題四之網格法結果 2D 圖

表 15 例題五之網格法結果

核心係數 g

懲罰係數C 0.0001 0.001 0.01 0.1 1 10 100 1000

0.01

16.1% 16.1% 16.1% 43.6% 52.8% 26.1% 16.1% 16.1%

0.1

16.1% 16.1% 46.5% 54.8% 64.3% 72.1% 23.3% 16.1%

1

16.1% 46.4% 54.9% 63.5% 71.5% 77.5% 78.7% 48.2%

10

46.3% 54.9% 63.3% 69.3% 76.9% 79.1% 80.0% 49.6%

100

54.9% 63.1% 67.7% 72.5% 78.1% 80.5% 79.3% 49.6%

1000

63.1% 67.3% 69.6% 76.5% 78.5% 80.2% 78.8% 49.6%

10000

66.7% 68.7% 71.9% 77.4% 78.2% 79.7% 78.6% 49.6%

100000

68.0% 69.3% 74.7% 78.0% 77.9% 78.8% 78.6% 49.6%

0.0 001 0. 001 0.01 0.1 1 10 100 1000

0.01 0.1 1 10 100 1000 10000 100000

核心係數g

懲罰係 數 C

80.0%-85.0%

75.0%-80.0%

70.0%-75.0%

65.0%-70.0%

60.0%-65.0%

55.0%-60.0%

50.0%-55.0%

圖 5 例題五之網格法結果 2D 圖

表 16 例題六之網格法結果

核心係數 g

懲罰係數C 0.0001 0.001 0.01 0.1 1 10 100 1000

0.01

0.0111 0.0109 0.0092 0.0054 0.0049 0.0075 0.0105 0.0107

0.1

0.0109 0.0092 0.0054 0.0041 0.0040 0.0058 0.0093 0.0096

1

0.0091 0.0055 0.0049 0.0033 0.0035 0.0058 0.0096 0.0100

10

0.0055 0.0055 0.0042 0.0029 0.0037 0.0058 0.0096 0.0100

100

0.0056 0.0051 0.0040 0.0026 0.0037 0.0058 0.0096 0.0100

1000

0.0057 0.0042 0.0033 0.0028 0.0042 0.0058 0.0096 0.0100

10000

0.0052 0.0041 0.0027 0.0057 0.0042 0.0058 0.0096 0.0100

100000

0.0044 0.0036 0.0029 0.0076 0.0042 0.0058 0.0096 0.0100

0.0 001 0. 001 0.01 0.1 1 10 100 1000

0.01 0.1 1 10 100 1000 10000 100000

核心係數g

懲罰係 數 C

0.011 -0.012 0.010 -0.011 0.009 -0.010 0.008 -0.009 0.007 -0.008 0.006 -0.007 0.005 -0.006 0.004 -0.005 0.003 -0.004 0.002 -0.003

圖 6 例題六之網格法結果 2D 圖

表 17 例題七之網格法結果

核心係數 g

懲罰係數C 0.0001 0.001 0.01 0.1 1 10 100 1000 0.01 49.6% 49.6% 49.6% 67.9% 76.0% 49.6% 49.6% 49.6%

0.1 49.6% 49.6% 68.9% 74.8% 79.0% 80.3% 49.6% 49.6%

1 49.6% 69.1% 73.1% 77.3% 80.4% 80.9% 67.0% 53.0%

10 69.4% 72.8% 76.0% 78.6% 80.3% 80.1% 67.9% 53.8%

100 73.0% 75.0% 77.0% 79.9% 78.3% 78.3% 67.9% 53.8%

1000 74.0% 75.9% 78.0% 79.8% 76.3% 78.3% 67.9% 53.8%

10000 74.5% 76.9% 78.8% 79.6% 75.0% 78.3% 67.9% 53.8%

100000 75.9% 78.0% 79.1% 79.0% 75.8% 78.3% 67.9% 53.8%

0.0 001 0. 001 0.01 0.1 1 10 100 1000

0.01 0.1 1 10 100 1000 10000 100000

核心係數g

懲罰係 數 C

80.0%-85.0%

75.0%-80.0%

70.0%-75.0%

65.0%-70.0%

60.0%-65.0%

55.0%-60.0%

50.0%-55.0%

圖 7 例題七之網格法結果 2D 圖

表 18 例題八之網格法結果

核心係數 g

懲罰係數C 0.0001 0.001 0.01 0.1 1 10 100 1000 0.01 7.0056 6.9511 6.6336 6.7580 7.0102 7.0128 7.0128 7.0128

0.1 6.9482 6.4630 4.5659 5.2467 6.9945 7.0206 7.0206 7.0206 1 6.4401 3.9433 3.0225 3.7298 6.8012 7.0284 7.0284 7.0284 10 3.8925 3.0146 3.2230 4.1622 6.5416 7.0680 7.0680 7.0680 100 3.0274 3.1749 3.7783 7.7081 6.5416 7.0680 7.0680 7.0680 1000 3.1361 3.2545 4.4794 9.0668 6.5416 7.0680 7.0680 7.0680 10000 3.2457 3.6059 8.7739 9.0668 6.5416 7.0680 7.0680 7.0680 100000 3.3049 4.1269 24.9572 9.0668 6.5416 7.0680 7.0680 7.0680

0.0 001 0. 001 0.01 0.1 1 10 100 1000

0.01 0.1 1 10 100 1000 10000 100000

核心係數g

懲罰係 數 C

9.0 -10.0 8.0 -9.0 7.0 -8.0 6.0 -7.0 5.0 -6.0 4.0 -5.0 3.0 -4.0

圖 8 例題八之網格法結果 2D 圖

表 19 例題九之網格法結果

核心係數 g

懲罰係數C 0.0001 0.001 0.01 0.1 1 10 100 1000 0.01 55.6% 55.6% 55.6% 55.6% 55.6% 55.6% 55.6% 55.6%

0.1 55.6% 55.6% 63.3% 83.3% 61.1% 55.6% 55.6% 55.6%

1 55.6% 67.0% 83.7% 83.7% 83.3% 63.0% 55.6% 54.4%

10 67.4% 84.1% 83.7% 80.0% 81.5% 63.3% 55.2% 54.4%

100 84.1% 83.7% 84.4% 80.4% 79.3% 63.7% 55.2% 54.4%

1000 83.7% 84.1% 79.6% 77.4% 79.6% 62.6% 55.2% 54.4%

10000 84.8% 83.7% 78.9% 77.0% 78.1% 62.2% 55.2% 54.4%

100000 83.7% 79.6% 79.3% 78.5% 72.2% 62.2% 55.2% 54.4%

0.0 001 0. 001 0.01 0.1 1 10 100 1000

0.01 0.1 1 10 100 1000 10000 100000

核心係數g

懲罰係 數 C

80.0%-85.0%

75.0%-80.0%

70.0%-75.0%

65.0%-70.0%

60.0%-65.0%

55.0%-60.0%

50.0%-55.0%

圖 9 例題九之網格法結果 2D 圖

在文檔中 中 華 大 學 (頁 64-90)