臺灣博碩士論文加值系統

已有許多模糊模型被提出來用以評估學生成績表現，然而，這些方法都仰賴專家的主觀意見，沒有辦法建立客觀的歸屬函數。為了解決上述的問題，本研究利用CPDA(累積機率分配法)化模糊規則來評估學生成績等第。提出的模型有三個目標：(1)運用CPDA來產生屬性的歸屬函數值以得到更主觀的歸屬值、(2)採用不同的相似度函數門檻來產生規則，使提出的方法能夠以更適合的規則來處理分類問題、(3)運用相似度當作權重來產生模糊規則，藉此提高分類的品質、(4)運用粗糙集理論簡化產生規則的複雜度，並增加分類問題的效能。論文提出了兩個研究模型，研究模型A產生少量的規則但是仍能維持校能。運用模糊理論的研究模型B，用比較不複雜的方式產生模糊規則，且擁有的效能高過所有比較的模型。SPA50A和SPA50B資料集是實驗中的訓練資料集，15筆額外給予的資料當做測試資料集。提出的模型A產生的規則數跟陳提出的模型產生的規則數一樣，都是五條，這樣的規則數目讓人能夠比較容易去產是規則的意義，而且模型A的效能也很好。模型B運用粗糙集理論來產生規則，並且獲得所有比較的模型中最好的效能。運用了粗糙級理論的模型B，產生規則的方式比較簡單，而且又能夠擁有所有比較的模型中最佳的性能。本研究中最重要部分就是找出老師在經過判斷後，對學生給定成績的模式，透過所提出方法產生的規則，我們可以觀察到老師給定成績的習性和模式。

Several fuzzy models have been proposed to evaluate student academic performance in the past. However, these models are based on subjective expert’s opinions, so they don’t build objective membership function grade. To overcome the problem above, this study utilizes CPDA (Cumulative Probability Distribution Approach) based FRFE (Fuzzy Rule from examples) model to evaluate students’ grade, the proposed models has three objectives: (1) utilize CPDA to generate membership function grade of attributes in order to obtain more objective membership function grade; (2) by adopting different similarity function thresholds to generate rules, the proposed method generates more fit rules to solve classification problems; (3) use similarity as weight to generate fuzzy rules in order to improve the quality of classification, and (4) by adopting rough set theory, simplify the complexity of generating rule and increases the performance in classification problems. Two research models are proposed. Research model A has few rules without losing good performance. By adopting rough set theory, research model B proposes a simple way to generate fuzzy rules and has superior performance than the listing models have. SPA50A and SPA50B dataset are used for training in experiments. Another 15 given records are used for testing. Proposed model A has good performance and the same number of rules as Chen’s model has. Five rules are generated in proposed model A, and because the number of rules is few, they are easier to be interpreted by human. Proposed model B uses rough set theory to generate rules and outperforms all listing models. By adopting rough set theory, proposed model B generate rules in a simpler way and still has the best performance among all listing models. Finding patterns of how teachers assign grades to students by their judgments is the most important part in this study. By the rules that proposed models generated, we can observe the habits and patterns of how teachers assigning grades to students.

摘要 i
ABSTRACT ii
誌謝 iv
Content v
List of Figures vi
List of Tables vii
1. Introduction 1
1.1 Background 1
1.2 Motivation 2
1.3 Objective 3
1.4 Research Limitations 4
1.5 Thesis Organization 5
2. Related Works 6
2.1 Fuzzy Set Theory 6
2.2 Similarity Function 7
2.3 Cumulative Probability Distribution Approach (CPDA) 9
2.4 Generation of Fuzzy Rules 11
2.4.1 Yuan and Shaw’s model 11
2.4.2 Chen et al’s model 12
2.4.3 Biswas’ model 13
2.4.4 Chen and Lee’s model 14
2.4.5 NEFCLASS model 15
2.5 Evaluation of Student Grade 16
2.6 Rough set theory 18
3. Research Methodology 21
3.1 Research Model A 22
3.1.1 Concept of research model A 23
3.1.2 Proposed algorithm of research model A 25
3.2 Research Model B 34
3.2.1 Concept of research model B 34
3.2.2 Proposed algorithm of research model B 36
4. Experimental Results and Comparison 40
4.1 Research model A verification 41
4.2 Results of research model B verification 44
4.3 Comparison 47
4.4 Findings 49
5. Conclusion 50
References 51

1.Biswas R., “An application of fuzzy sets in students’ evaluation”, Fuzzy Sets Syst 74:187–194, 1995.
2.Castro J.L. and Zurita J.M., “An inductive learning algorithm in fuzzy systems”, Fuzzy Sets and Systems 89(2):193± 203, 1997.
3.Chen S.M. and Lee C.H., “New methods for students’ evaluation using fuzzy sets”, Fuzzy Sets Syst 104:209–218, 1999.
4.Chen S.M., Lee S. H. and Lee C.H, “Generating fuzzy rules from numerical data for handling fuzzy classification problems”, Proceedings of the 1999 National Computer Symposium, Taipei, Taiwan, Republic of China, Vol. 2, 336± 343, 1999.
5.Chen S.M., Lee S.H. and Lee C.H., “A new method for generating fuzzy rules from numerical data for handling classification problems”, 2001.
6.Chen S.M. and Yeh M.S, “Generating fuzzy rules from relational database systems for estimating null values”, Cybernetics and Systems: An International Journal 29(6):363± 376, 1998.
7.Christensen, R., “Entropy minimax sourcebook”, Vol. 1-4, and “Fundamentals of inductive reasoning”, Entropy Ltd., Lincoln, MA, 1980.

8.Dallal G.E. and Wilkinson, L., “An analytic approximation to the distribution of Lilliefors'' test for normality”, The American Statistician, Vol. 40, pp. 294–296, 1986.
9.Gipps C., Stobart G., “Assessment: A teachers’ guide to the issue”, Hodder and Stoughton, London, 1993.
10.Goldstein H., “Statistical and psychometric models for assessment”, In: Goldstein H, Lewis T (eds) Assessment: Problems, developments and statistical issues, JohnWiley&Sons, Chichester. pp 41–55, 1996.
11.Hayashi Y., and Imura A., “Fuzzy neural expert system with automated extraction of fuzzy if-then rules from a trained neural network”, Proceedings of the 1990 First International Symposium on Uncertainty Modeling and Analysis, University of Maryland, College Park, MD, 1990.
12.Holte R.C. “Very simple classification rules perform well on most commonly used datasets”, Machine Learning 11:63-91, 1993.
13.Hong T.P., and Lee C.Y., “Induction of fuzzy rules and membership functions from training examples”, Fuzzy Sets and Systems 84(1):33± 47, 1996.
14.Klawonn F., and Kruse R., “Constructing a fuzzy controller from data”, Fuzzy Sets and Systems 85(2):177± 193, 1997.

15.Kosko B., “Fuzzy entropy and conditioning”, Information Sciences 40(2):165-174, 1986.
16.Nauck D., “Measuring interpretability in rule-based classification systems”, Proceedings of the International Conference on Fuzzy Systems, St. Louis, Missouri, 2003.
17.Nozaki K., Ishibuchi H., and Tanaka H., “A simple but powerful heuristic method for generating fuzzy rules from numerical data”, Fuzzy Sets and Systems 86(3):251± 270, 1997.
18.Pawlak Z., “Rough sets”, Int. J. Computer and Infn. Sci., 11(5), 341–356, 1982.
19.Pawlak Z. and Skoworn A., "Rudiments of rough sets", InformationScience 177, 2007.
20.Quinlan J.R., C4.5: Programs for Machine Learning, Morgan Kaufmann Publishers, San Mateo, CA.
21.Rasmani and Shen, “Data-driven fuzzy rule generation and its application for student academic performance evaluation”, 2006.
22.Ross, T.J., “Fuzzy logic with engineering applications, International edition”, McGraw-Hill, USA, 2000.
23.Ross T.J., “Fuzzy logic with engineering applications, Second edition”, John Wiley, USA, 2004.
24.Wang L.X., and Mendel J.M., “Generating fuzzy rules by learning from examples”, IEEE Transactions on Systems, Man, and Cybernetics 22(6):1414± 1427, 1992.
25.Wong J.T. and Y.S. Chung, "Rough set approach for accident chains exploration", Accident Analysis and Prevention 39: 629-637, 2007.
26.Wu T.P., and Chen S.M., “A new method for constructing membership functions and fuzzy rules from training examples”, IEEE Transactions on Systems, Man, and Cybernetics ± Part B 29(1):25± 40, 1999.
27.Yager R. and Filev, D., “Template-based fuzzy system modeling”, Intelligent and Fuzzy Sys., Vol. 2, pp. 39-54, 1994.
28.Yuan Y. and Shaw M.J., “Induction of fuzzy decision trees”, Fuzzy Sets and Systems 69(2):125-139, 1995.
29.Yuan Y., and Zhuang H., “A genetic algorithm for generating fuzzy classification rules”, Fuzzy Sets and Systems 84(1):1± 19, 1996.
30.Zadeh L.A., “Fuzzy sets”, Information and Control 8:338-353, L. A., 1965.