第五章 結論與建議
第二節 後續相關研究建議
傳統迴歸模式已被廣泛地使用,且統計迴歸理論至今已發展的相當完備,然 而有些問題卻是傳統迴歸模式難以處理的。當欲處理的資料具共線性時,便無法 準確地處理,一般改善方法為採用脊迴歸模式,本研究採用以模糊積分與模糊支 撐為基礎之Choquet 積分迴歸模式來進行處理。且模糊測度及模糊積分的概念已被 廣泛地應用在各種學科領域中,依其所應用的情形做適當之改變,但常用之測度 是以Sugeno 之 λ 測度及 Zadeh 之 P 測度為主,在上述兩種測度不符合實際需求時,
經由本研究分析比較結果,建議可採用基於混合
L
δ-模糊測度與γ -支撐之 Choquet 積分迴歸模式進行研究。根據文獻探討及研究結果,提出以下幾點建議,並說明未來可能的研究方向,
以供參考與發展:
一、發展新的模糊測度並應用在基於γ -模糊支撐的Choquet積分迴歸模式。
二、模擬實驗:本研究因為時間上的限制,所以未進行模擬實驗,在後續的研究 中,可再多增加人數,進行模擬實驗研究。
三、本研究所採用之教育測驗資料,所選取之樣本資料對於自變數與應變數之間 的γ 支撐呈現正值,若日後使用其餘實證資料或模擬資料時,γ 支撐呈現負值 時需再加以模式轉換,才能增進預測效果。
42
四、本研究是比較基於γ -模糊支撐與模糊測度之 Choquet 積分迴歸模式與複線性 迴歸模式、脊迴歸模式之預測效力,在後續研究中可將Choquet 積分廣泛應用 於IRT 模式、GHMM 模式並與無參數結合,甚至廣泛應用於企業管理、財務 金融等方面。
43
參考文獻
國中基本學力測驗推動工作委員會(2007)。九十六年國民中學基本學力測驗Q &
A。取自:http://www.bctest.ntnu.edu.tw/96QA.htm
劉湘川(2006a)。基於P測度之改進模糊測度及其模糊積分。測驗統計年刊,14
(上),1-15。臺中市:國立臺中教育大學。
劉湘川(2006b)。λ 測度之改進模糊測度及其模糊積分。測驗統計年刊,14(上), 16-34。臺中市:國立臺中教育大學。
劉應興(1997)。應用線性迴歸模型。台北市:華泰書局。
Browne, M. W. (2000). Cross-validation methods. Journals of Mathematical Psychology, 44, 108-132.
Choquet, G. (1953). Theory of capacities. Annales de l’Institut Fourier, 5, 131-295.
Dempster, A. P. (1967). Upper and lower probabilities induced by multi-valued mapping.
Annals of Mathematical Statistics, 38, 325-339.
Hoerl, A. E., & Kennard, R. W. (1970). Ridge regression: Applications to nonorthogonal problems. Technometrics, 12, 69-82.
Hoerl, A. E., Kennard, R. W., & Baldwin, K. F. (1975). Ridge regression: Some simulation. Communications in Statistics, 4, 105-123.
Liu, H. C., Chen, C. C., Chen, G. S., & Jheng, Y. D. (2007). Power-transformed-measure and its Choquet Integral Regression Model. Proceedings of WSEAS International Conference on Applied Computer Science (ACS’07), 7, 101-104.
Liu, H. C., Chen, C. C., Jheng, Y. C., & Chien, M. F. (2009). Choquet integral with respect to extensional L-measure and its application. Proceedings of International Conference on Fuzzy Systems and Knowledge Discovery, 6, 131-136.
Liu, H. C., Chen, C. C., Wu, D. B., & Sheu, T. W. (2009). Theory and Application of the Composed Fuzzy Measure of L-Measure and Delta-Measures, Journals of WSEAS Transactions on Information Science and Control, 8(4), 359-368.
Liu, H. C., Tu, Y. C., Chen, C. C., & Wang, W. S. (2008). The Choquet integral with respect to λ-measure based on γ-support. Proceedings of International Conferences
44
on Machine Learning and Cybernetics, 7, 3602-3606.
Liu, H. C., Tu, Y. C., Huang, W. C., & Chen, C. C. (2008). The Choquet Integral with Respect to R-Measure Based on γ-Support. Proceedings of The 5th International Conference on
Fuzzy Systems and Knowledge Discovery (FSKD 2008), doi: 10.1109/FSKD.2008.545,
645649.Liu, H. C., Tu, Y. C., Lin, W. C., & Chen, C. C. (2008). Choquet integral regression model based on L-measure and γ -support. Proceedings of the 2008 International Conference on Wavelet Analysis and Pattern Recognition, 777-782.
Liu, H. C., Wu, D. B., Jheng, Y. D., & Sheu, T. W. (2009). Theory of Multivalent Delta-Fuzzy Measures and its Application. Journals of WSEAS Transactions on Information Science and Application, 6(6), 1061-1070.
Shafer, G. (1976). A Mathematical Theory of Evidence. Princeton, NJ: Princeton University.
Sugeno, M. (1974). Theory of fuzzy integrals and its applications. Unpublished Ph.D.
thesis, Tokyo Institute of Technology, Tokyo, Japan.
Weiner, J. S., & Lourie, J. A. (1981). Practical Human Biology. London: Academic Press.
Zadeh, L. A. (1978). Fuzzy sets as a basis for a theory of possibility. Fuzzy Sets and Systems, 1, 3-28.
45
附錄
附錄一 數學、理化、生物、地球科學畢業成績及其國中基本學力測驗自然科成績
No. C1 C2 C3 C4 BCT No. C1 C2 C3 C4 BCT
1 72 66 78 72 19 31 66 68 75 74 25
2 86 80 82 81 35 32 68 70 74 76 40
3 56 63 69 75 21 33 57 65 75 70 24
4 78 86 86 86 33 34 74 70 80 75 35
5 66 72 80 76 23 35 49 60 69 64 13
6 68 74 77 80 28 36 51 60 63 64 18
7 74 86 87 88 44 37 58 64 68 66 32
8 54 56 62 68 7 38 73 78 84 81 39
9 71 74 80 77 26 39 56 56 65 61 6
10 68 70 80 75 33 40 61 62 70 70 25
11 53 56 70 63 22 41 57 60 68 64 23
12 67 70 80 75 35 42 57 64 67 70 26
13 70 66 70 74 13 43 50 52 68 60 7
14 60 65 75 70 23 44 84 80 76 72 49
15 68 68 78 76 35 45 62 66 76 71 22
16 58 66 76 71 37 46 70 74 78 82 32
17 61 66 72 78 33 47 69 70 80 75 26
18 68 68 80 74 26 48 63 74 74 74 42
19 56 66 76 71 21 49 66 78 80 82 39
20 59 62 70 78 29 50 67 70 80 75 31
21 62 64 76 70 36 51 56 65 75 70 23
22 71 72 78 75 26 52 50 54 66 60 18
23 74 63 69 75 12 53 71 75 85 80 41
24 59 70 80 76 37 54 74 77 80 85 26
25 75 75 85 80 39 55 71 72 76 80 31
26 73 78 84 81 24 56 60 65 75 70 21
27 62 68 72 74 29 57 59 57 70 68 17
28 77 74 80 76 42 58 50 56 65 68 13
29 63 60 68 69 17 59 72 76 80 78 38
30 56 61 75 68 22 60 81 76 78 80 33 C1:理化
C2:生物 C3:地球科學 C4:數學
BCT:自然科成績
46
附錄二 128 位學生體胖百分比、二頭肌、三頭肌、肩胛骨下肌及腸胃上肌測定值 No BIA A1 A2 A3 A4 No BIA A1 A2 A3 A4
1 11.8 15.8 16.2 27.6 16.3 65 22.8 27.4 32.2 33.6 24.5 2 17.8 20.4 12.8 35.6 23.0 66 5.4 14.4 8.0 17.2 17.5 3 6.2 10.4 9.4 12.6 8.2 67 8.6 9.6 11.4 14.4 14.8 4 7.0 10.2 5.6 10.6 13.6 68 7.4 9.6 8.0 10.2 13.4 5 5.2 10.8 12.0 13.8 11.8 69 20.2 29.2 28.2 40.0 23.9 6 20.2 24.6 39.8 40.0 34.7 70 28.4 30.6 38.6 39.8 39.5 7 14.8 18.8 22.4 24.8 18.6 71 6.8 10.4 8.8 14.6 15.2 8 12.4 18.2 20.2 27.6 18.4 72 11.0 12.4 10.4 14.4 11.7 9 17.0 20.6 21.2 27.8 21.6 73 20.8 30.8 38.8 40.0 37.7 10 16.6 22.6 18.8 32.6 16.6 74 25.0 27.6 40.0 40.0 29.6 11 10.2 13.4 8.4 12.8 15.1 75 4.8 6.2 5.6 7.0 9.7 12 21.0 26.7 36.2 40.0 23.9 76 8.4 7.0 8.0 11.8 15.6 13 10.0 13.4 17.4 17.8 18.1 77 11.0 14.0 10.0 12.4 17.7 14 9.8 13.8 11.6 19.2 15.5 78 10.6 11.6 7.0 14.4 12.9 15 8.6 10.8 11.4 14.4 8.5 79 12.4 16.8 17.0 25.8 14.9 16 21.8 25.6 36.6 36.8 31.7 80 11.4 14.0 20.8 26.4 25.1 17 25.2 29.4 30.2 30.6 31.2 81 12.8 15.8 20.2 19.8 16.4 18 10.2 11.0 10.2 14.8 12.2 82 12.0 15.6 16.0 22.8 17.3 19 10.4 15.8 13.0 25.6 16.6 83 13.0 20.4 24.0 27.0 26.2 20 17.6 22.6 23.2 34.6 23.2 84 14.2 15.4 22.4 22.8 18.9 21 12.6 12.4 14.2 16.0 14.2 85 11.0 16.4 14.2 15.8 17.9 22 9.0 11.2 9.4 11.8 9.3 86 22.4 29.8 35.0 36.2 28.5 23 12.2 19.2 17.4 27.8 19.7 87 6.4 7.6 8.6 11.4 12.1 24 4.6 7.0 8.8 11.2 7.2 88 6.8 10.6 9.6 14.6 12.2 25 6.4 8.8 11.0 12.6 10.8 89 16.2 18.4 27.2 27.4 24.9 26 23.8 29.0 37.0 35.0 30.7 90 22.4 26.8 25.4 33.4 30.4 27 8.4 15.8 17.8 23.0 21.6 91 9.6 11.2 10.4 18.0 11.7 28 12.2 16.6 16.4 20.6 18.7 92 10.8 17.2 24.0 24.8 21.9
47
No BIA A1 A2 A3 A4 No BIA A1 A2 A3 A4 29 7.2 12.8 8.6 18.6 15.2 93 13.0 16.2 12.4 18.4 14.2 30 21.4 31.2 31.4 39.4 28.6 94 5.6 12.4 11.4 15.6 14.5 31 18.2 23.0 40.0 40.0 28.2 95 19.4 25.0 36.2 39.0 29.9 32 9.2 12.6 40.0 17.8 16.0 96 14.4 22.4 29.8 35.0 24.8 33 10.2 18.8 17.8 20.8 18.4 97 25.4 29.4 37.0 40.0 24.6 34 19.2 24.4 35.2 35.0 34.1 98 9.4 11.2 11.4 12.4 8.9 35 6.8 12.0 8.0 14.4 16.1 99 17.4 22.6 19.4 31.6 22.7 36 16.8 20.8 25.6 27.8 20.7 100 24.0 30.8 40.0 40.0 29.4 37 35.8 38.6 40.0 40.0 30.1 101 3.8 6.0 6.4 6.8 10.8 38 10.0 11.6 10.4 18.6 8.3 102 11.0 19.4 11.6 18.4 13.7 39 5.4 12.2 12.4 21.4 19.2 103 22.6 24.4 40.0 40.0 33.3 40 11.2 18.0 23.6 30.8 22.1 104 9.2 10.0 11.0 19.2 13.4 41 5.4 11.2 6.8 11.6 11.9 105 18.2 19.0 31.0 29.4 24.5 42 7.6 8.4 9.4 13.6 12.8 106 6.8 12.4 14.0 17.8 14.1 43 6.6 9.8 9.6 12.0 9.3 107 7.4 11.6 10.0 16.0 11.0 44 32.4 37.2 40.0 40.0 18.2 108 9.2 10.6 12.4 14.4 12.7 45 7.8 14.0 11.0 17.8 31.3 109 29.4 23.6 39.8 40.0 37.4 46 17.8 26.6 34.2 40.0 22.8 110 6.8 7.8 9.8 12.8 10.9 47 22.0 27.8 38.2 39.4 23.6 111 12.4 14.6 15.8 19.8 16.9 48 14.4 15.8 18.8 23.8 14.9 112 8.2 9.8 9.2 16.0 14.5 49 15.8 18.4 21.4 24.0 24.9 113 16.4 20.8 25.2 30.4 24.9 50 7.4 12.8 10.2 17.0 14.3 114 9.4 11.4 12.0 21.8 14.3 51 16.2 29.0 21.6 29.8 24.1 115 16.4 22.4 33.2 36.8 25.1 52 6.0 7.4 7.6 9.8 8.6 116 7.0 11.4 13.8 17.4 11.2 53 12.2 15.4 16.2 18.8 17.8 117 10.4 12.6 14.8 23.8 18.0 54 11.6 12.0 9.8 13.0 8.9 118 5.6 8.2 10.2 8.6 7.7 55 17.8 22.6 38.0 31.0 24.8 119 10.8 11.8 17.8 21.2 19.9 56 13.2 16.8 18.6 23.4 20.7 120 9.6 15.8 14.4 19.4 18.6 57 4.4 7.2 8.2 9.8 14.3 121 5.0 6.8 7.4 9.4 6.0
48
No BIA A1 A2 A3 A4 No BIA A1 A2 A3 A4 58 16.2 21.8 28.2 32.6 27.2 122 9.8 12.2 12.4 15.4 13.5 59 11.4 19.4 28.8 32.8 22.3 123 13.8 18.0 16.4 21.0 19.3 60 11.2 13.0 18.8 22.6 21.9 124 8.8 12.8 9.8 11.8 13.3 61 8.6 11.4 7.2 10.2 7.5 125 15.8 21.0 35.4 39.8 27.3 61 20.4 26.2 31.0 32.8 25.8 126 10.8 16.6 15.6 23.2 16.5 63 7.0 8.8 11.6 9.4 12.0 127 9.0 10.6 10.0 16.8 11.9 64 14.6 17.4 12.8 16.8 14.7 128 8.8 12.4 10.0 10.8 11.3
A1:二頭肌 A2:三頭肌 A3:肩胛骨下肌 A4:腸胃上肌 BIA:體胖百分比