在一個工程開始前，成本預估是各個廠商所要面對的重要議題。其中一項重要的任務是預測材料的未來價格以便控制庫存成本。支持向量回歸(support vector regression, SVR)是一個適合的預測工具，但是其對於非穩態的價格資料並沒有很好的預測效果，因此，探討出一個有效可行的預測方法還是一項巨大的挑戰。
在本論文中，我們以世界上主要的金屬材料「銅」為預測以及研究對象。預測方面，使用最小平方支持向量回歸(least squares support vector machine, LS-SVR)預測銅的未來趨勢和價格，並結合經驗模態分析(empirical mode decomposition, EMD)處理原始價格資料，經驗模態分析將非現性和非穩態的資料拆解為數個本質樣態函數(intrinsic mode function, IMF)和一個剩餘量。因此，結合所有透過在線最小平方支持向量回歸(online least squares support vector machine, online LS-SVR)來預測的IMF可以得到較好的預測結果。
除此之外，本論文透過EMD拆解出來的IMF和剩餘量所表現的中長期趨勢探討銅、黃金、白金、白銀、西德州原油、英鎊兌美元匯率、銅庫存以及中國與美國的經濟成長率之間的關係。

The cost estimate is an important issue faced by all manufacturers before starting off a project. The major task is to predict the future price of the material in order to control inventory costs. Despite the prediction system based on support vector machine (SVM) has recently become a good solution, the accuracy of the predicted values are usually deteriorated with non-stationary price data. Therefore, to further explore an effective and feasible method of price prediction is still a huge challenge in materials’ cost control.
In this thesis, least squares support vector regression (LS-SVR) is proposed to predict the trend of copper price. The design combines empirical mode decomposition (EMD) to decompose nonlinear and non-stationary data into several intrinsic mode function (IMF) components and one residual component. Therefore, better prediction can be attained by forecasting these IMFs and residual value individually with corresponding online LS-SVR. According to our test results, proposed design improves prediction accuracy from online LS-SVM for the trend of copper price.
In addition, we discuss the relationship between copper, gold, platinum, silver, West Texas Intermediate (WTI), GBP/USD exchange rate and inventory by the long-term trend of the sum of few IMFs and residue.

[1] H. J. Kim, Y. C. Seo and C. T. Hyun, “A hybrid conceptual cost estimating model for large building projects,” Automation in Construction, 25:72-81,2012.
[2] M. Y. Cheng, N. D. Hoang, T. T. Chong, “Hybrid intelligence approach based on LS-SVM and Differential Evolution for construction cost index estimation: A Taiwan case study,” Automation in Construction, 23:306-313, 2013.
[3] V. Vapnik, The nature of statistical learning theory, Springer, New York, 1995.
[4] H. Drucker, C. J. C. Burges, L. Kaufman, A. Smola and V. Vapnik, “Support vector regression machines,” Advanced in Neural Information Processing System, 9:155-161, 1997.
[5] J. A. K. Suykens and J. Vandewalle, “Least squares support vector machine classifiers,” Neural Processing Letters, 9:293-300, 1999.
[6] H. Wang and D. Hu, “Comparison of SVM and LSSVM for regression,” In Proceedings of the International Conference on Neural Networks and Brain (ICNN&B ’05), pages 279-283, Beijing, China, 2005.
[7] M. W. Chang, C. J. Lin and R. C. Weng, “Analysis of non-stationary time series using support vector machines,” Pattern Recognition with Support Vector Machines Lecture Notes in Computer Science, Springer, 2002.
[8] M. Zhang, K. Li and Y. Hu, “Classification of power quality disturbances using wavelet packet energy entropy and LS-SVM,” Energy and Power Engineering, 2:154-160, 2010.
[9] Fan, J and Tang, Y. “An EMD-SVR method for nonstationary time series prediction,” In proceedings of the International Conference on Quality, Reliability, Maintenance, and Safety Engineering, pages 1765-1770, Chengdu, China, 2013.
[10] C. S. Lin, S. H. Chiu and T. Y. Lin, “Empirical mode decomposition–based least squares support vector regression for foreign exchange rate forecasting,” Economic Modelling, 29:2583-2590, 2012.
[11] C. M. Vong, P. K. Wong, K. Li and R. Zhang, “A study on online LS-SVM for modelling of electronically-controlled automotive engine performance,” In proceedings of the International Symposium on Advanced Vehicle Control, Kobe, Japan, 2008. (6 pages)
[12] S. Chao, J. Zhang, F. Liu and M. Li, An online LS-SVM prediction model of building space cooling load. Advanced Materials Research, 354-355:789-793, 2012.
[13] N. E. Huang, Z. Shen and S. R. Long, “The empirical mode decomposition and the Hilbert spectrum for nonlinear and non-stationary time series analysis,” Proceedings of the Royal Society A: Mathematical, Physical and Engineering Sciences 454:903–995, 1998.
[14] M. Martin, “On-line support vector machine regression,” In proceedings of the 13th European Conference on Machine Learning, pages 282-294, Helsinki, Finland, 2002.
[15] Z. Dong, X. Tian and J. Zeng, “Mechanical fault diagnosis based on LMD approximate entropy and LSSVM,” TELKOMNIKA, 11(2):803-808, 2013.
[16] N. E. Huang, M.C. Wu and S. R. Long, “A confidence limit for the empirical mode decomposition and Hilbert spectral analysis,” Proceedings of the Royal Society A: Mathematical, Physical and Engineering Sciences 459:2317–2345, 2003.