在真實生活中，欲進行資料分析時，常囿於資料取得困難、時間、或成本的考量，無法取得有效數量之樣本，因此如何從所獲得之少量資料中，發掘有意義的資訊以建立穩定的知識模式，在近年來已是重要的議題。目前小樣本學習方法，主要可從資料的轉換、虛擬資料量之擴充、以及新屬性的開發等方面著手。在虛擬資料量擴充部份，本研究提出一個兩階段的虛擬樣本產生方法，第一階段使用模式樹中的M5'演算法對小樣本進行學習，依其分類法則產生各葉部節點之可能值；第二階段則依此可能值於各葉部節點產生應存在於小樣本中之樣本，稱之潛在樣本。實驗結果顯示，透過此方法產生之潛在樣本，在加入原小樣本後所建構之預測模式，確實較原小樣本對未知母體的預測誤差有明顯的改善，並且於方法比較上，本研究比整體趨勢擴散方法有更佳之效果。此外，本研究之主要貢獻在於小樣本之名目屬性資料之學習，冀能成為後續相關研究之參考。

In real world, while we try to learn information from data with machine learning algorithms, sometimes the size of acquirable data is quite small so that the learning results may not be reliable since the learning tools have their smallest required sample number respectively to ensure their robustness. The reasons of the small data learning problems occuring can be mainly owing to the consideration of time and cost. In this research, a two-phase procedure is proposed to solve such learning problems by generating more training samples for learning tools to improve their predictions. In the first phase, the M5’ model is employed to acqire classification rules which contain the corrleations between attributes and the possible values of each leaf-node can be thus obtained. In the second phase, samples store at a leaf-node can be permuted with correspondent possible values to generate the training samples which are very likely to happen are called potential samples. The experimental results reveal that the prediction accuracies can be improved while learning with potential samples, especially when the original sample size is small. Besides, compared with a virtual sample generation algorithm, Mega-Trend-Diffusion, the proposed procedure significantly outperforms.

葉怡成，“類神經網路模式應用與實作”，儒林圖書公司，民國92年3月。
Anthony, M., Biggs, N. (1997). Computational Learning Theory. Cambridge University Press.
Breiman, L., Friedman, J.H., Olshen, R.A., and Stone, C.J. (1984). Classification and Regression Trees, Belmont, CA：Wadsworth International Group.
Chaudhuri, P., Huang, M. C., Loh, W. Y.,& Rubin, R. (1994). Piecewise-polynomial regression tree, Statistica Sinica, Vol.4, 143-167.
Dobra, A. and Gehrke. J.E., (2002). SECRET： A Scalable Linear Regression Tree Algorithm, Proceedings of the Eighth ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, 481-487.
Drucker, H., Burges, C. J. C., Kaufman, L., Smola, A., and Vapnik, V. (1997). Support vector regression machines, in: M. Mozer, M. Jordan, and T. Petsche (Eds.), Advances in Neural Information Processing Systems, vol 9. Cambridge, MA: MIT Press, 1997, pp. 155-161.
Frank E., Wang Y., Inglis S., Holmes G., and Witten IH (1998). Technical note: Using model trees for classification. Machine Learning, 32, 63-76.
Huang, C. F., Moraga, C. (2004). A diffusion-neural-network for learning from small samples. International Journal of Approximate Reasoning, 35, 137-161.
Huang, C. J., Wang, H. F. (2009). Prediction of the Period of Psychotic Episode in Individual Schizophrenics by Simulation-Data Construction Approach.
Huang, C.F. (1997). Principle of information. Fuzzy Sets and Systems. 91,69-90.
Jang, J.S.R. (1993). ANFIS： adaptive-network-based fuzzy inference systems. IEEE Transactions on Systems, Man and Cybernetics. 23, 665-685.
Jennrich, R. I, Schluchter, M. D. (1986). Unbalanced repeated-measures models with structured covariance matrices. Biometrics, 42, 805-820.
Karalic, A. (1992). Employing linear regression in regression tree leaves, Proceeding of the 10th European Conference on Artificial Intelligence, 440-441.
Kerber, R. (1992). ChiMerge: discretization of numeric attributes, The Ninth ACM SIGKDDInternational Conference on Knowledge Discovery and Data Mining, Paris, France,123-128.
Laird, N. M., Ware, J. H. (1982). Random-effects models for longitudinal data. Biometrics, 38, 963-974.
Li, D. C., Chen, L. S., Lin, Y. S. (2003). Using functional virtual population as assistance to learn scheduling knowledge in dynamic manufacturing environments. International Journal of Production Research 41(17), 4011-4024.
Li, D. C., Hsu, H. C., Tsai, T. I., Lu, T. J., Hu, S. C. (2007a). A new method to help diagnose cancers for small sample size. Expert Systems with Applications, 33 420-424.
Li, D. C., Lin, Y. S. (2006). Using virtual sample generation to build up management knowledge in the early manufacturing stages. European Journal of Operational Research, 175 413-434.
Li, D. C., Tsai, T. I., Shi, S. (2009). A prediction of the dielectric constant of multi-layer ceramic capacitors using the mega-trend-diffusion technique in powder pilot runs: Case study. International Journal of Production Research 19, 51-69.
Li, D. C., Wu, C., Chen, F. M. (2005). Using Data-fuzzification Technology in Small Data Set Learning to Improve FMS Scheduling Accuracy. International Journal of Advanced Manufacturing Technology 27, 321-328.
Li, D. C., Wu, C., Tsai, T. I., and Chang, F. M. (2006). Using Mega-Fuzzification and Data Trend Estimation in Small Data Set Learning for Early FMS Scheduling Knowledge. Computers & Operations Research 33, 1857-1869.
Li, D. C., Wu, S. S., Tsai, T. I., Lina, Y. S. (2007b).Using mega-trend-diffusion and artificial samples in small data set learning for early flexible manufacturing system scheduling knowledge, Computers and Operations Research, 34, 966-982
Loh, W.Y. (2002). Regression trees with unbiased variable selection and interaction detection. Statistica Sinica, 12, 361-386.
Malerba, D., Esposito, F., Ceci, M.,& Appice, A. (2004). Top-Down Induction of Model Trees with Regression and Splitting Nodes. IEEE Transactions on Pattern Analysis and Machine Intelligence, Vol. 26, No. 5, 612-625.
Niyogi,P., Girosi, F., Tomaso, P. (1998). Incorporating prior information in machine learning by creating virtual examples. Proceeding of the IEEE 275-298.
Quinlan, J. R. (1979). “Discovering rules from large collections of examples:a case study.” In Michie, D., editor , Expert System in the Microelectronic Age. Edinburgh Scotland: Edinburgh University Press.
Quinlan, J. R. (1992). Learning with continuous classes. Proceedings of the Australian Joint Conference on Artificial Intelligence, 343-348.
Quinlan, J. R. (1993). “C4.5:Programs for Machine Learning.” San Mateo CA :Morgan Kaufmann.
Ribeiro, B. (2002). On Data Based Learning Using Support Vector Clustering. Proceeding of the 9th International Conference on Neural Information Processing (ICONIP’02), Vol. 5,2516-2521.
Roiger, R. J.,& Geatz, M. W. (2003). Data Mining：A Tutorial-Based Primer. New York：Addison Wesley.
Schervish, M. (1995). Probabilistic inference and inference diagrams. Operations Research 36, 589-604.
Vapnik, V. N. (1995). The Nature of Statistical Learning Theory, Springer-Verlag, New York.
Wang, H. F., Huang, C. J. (2008). Data construction method for the analysis of the spatial distribution of disastrous earthquakes in Taiwan. 189~212
Wang, Y. (2003). On-Demand Forecasting of Stock Prices Using a Real-Time Predictor. IEEE Transactions on Knowledge and Data Engineering. 1033-1037 Wang, Y., and Witten, I.H. (1997). Inducing model trees for continuous classes, Proceedings of poster papers of the 9th European Conference on Machine Learning.
Wang, Y., Song Q., MacDonell S., Shepperd M., Shen J. (2009). Integrate the GM(1,1) and Verhulst Models to Predict SoftWare Stage Effort. 647-658
Witten, I. H.,& Frank, E. (2005). Data Mining：Practical Machine Learning Tools and Techniques (2nd ed.). San Francisco：Morgan Kaufmann.