以往文獻指出執行價差交易並非像套利一樣無風險，因未來的價差走勢未必會如預期收斂，而造成損失，故本研究結合持有成本理論與預期理論於價差交易策略中，首先以持有成本理論建構無價差區間，發掘出可行價差時點，再以預期理論為基底，預測下一點價差走勢來決定進場時點以輔助投資者鎖定較佳的套利報酬，並減少風險存在。在預測工具的選擇上，則採用近年來於分類、預測領域表現傑出，卻尚未被應用於價差實證研究之平滑支撐向量機(SSVM)作為本研究之預測工具。
研究樣本取樣期間為2006年1月19日至2006年9月20日之日內五分鐘資料，以台股指數期貨(TX)、電子類股指數期貨(TE)、金融與保險類股指數期貨(TF)建構價差模型。經實證研究後，本研究發現：
一、各組價差模型仍具有相當多價差交易機會，以電子類股指數期貨和金融與保險類股指數期貨間價差交易機會最多。
二、預測工具方面，應用SSVR於價差模型上預測價差比率，高達八成以上準確率，且任一模型皆有不錯報酬表現，顯示SSVR適合應用於價差交易策略上。
三、在任一價差模型上，有使用預測工具其平均獲利報酬和勝率皆高於不使用預測工具，顯示價差擇時策略對投資者而言具有實質幫助的。
四、在考量風險觀念之後，有使用預測工具之風險溢酬皆高於不使用預測工具，且電子類股指數期貨和金融與保險類股指數期貨此組價差組合，擁有較高的風險溢酬。
五、輸入變數方面，本研究發現，在任何標的組合下，經過Pearson’s 係數檢定篩選出變數作為SSVR之輸入變數，比輸入所有變數之準確率提高約一成準確率。

In the literature, many studies address that spread is not risk free like arbitrage, because tendency of spread may not be expected to convergence and causes damage.
This research proposes the spread strategy which combines cost-of-carry model with expectation theory. First, the study applied cost-of-carry model to construct “no-spread-boundaries” and discovered the timing of spread. In addition, through the expectation theory, the next spread ratio tendency is predicted to determine the time of entering the market which can assist trader lock better than spread performance. As prediction tools, smooth support vector machines (SSVM) is adopted. SSVM outstands on classification and prediction in recent years, but has not been applied to spread studying yet.
The period of the sample selection is the intraday five-minute data from February 19, 2006 to September 20, 2006, By employing TX, TE and TF to construct the model of spread. The findings are as follows:
1. There are many opportunities of spread in every spread model, especially in the model of TE and TF.
2. In the aspect of prediction, the accuracy rate of SSVR is up to more than 80%, and every SSVR model has good return that shows SSVR is suitable for the spread strategy.
3. For any spread models, the average rate of return with predictive tools is higher than that without predictive tools. It is helpful for traders to apply the spread-timing-selection strategy.
4. After considering risk, the risk premium with predictive tools is higher than that without predictive tools, especially the model of TE and TF.
5. As input parameter, by assaying and screening the introduction parameters as the ones of SSVR through Pearson's coefficient, raising 10% about accuracy than all parameters of inputting in the association of any model.

1.王春笙 (1996)，「以技術指標預測台灣股市股價漲跌之實證研究—以類神經網路與複迴歸模式建構」，國立台灣大學資訊管理研究所碩士論文。
2.吳秉奇 (1999)，「類神經網路在台股指數期貨的預測與應用」，國立中央大學資訊管理研究所碩士論文。
3.李凌均 (2004)，「基於支持向量機的機械設備狀態趨勢預測研究」，西安交通大學學報，第38卷第3期，230-234頁。
4.李政彬 (2000)，「TAIFEX台股指數期貨與類股指數期貨價差交易之研究」，國立台灣大學財務金融學研究所碩士論文。
5.林建成 (2002)，「遺傳演化類神經網路於台灣股市預測與交易策略之研究」，私立東吳大學經濟學系碩士論文。
6.林政謙 (2005)，「應用平滑支撐向量機與類神經網路預測台灣上市電子指數漲跌之研究」，私立東吳大學經濟學系碩士論文。
7.徐俊明，「投資學理論與實務」，新陸書局。
8.陳昱升 (2006)，「台指期貨與摩台指期貨跨市場價差交易分析」，私立淡江大學財務金融學系碩士班。
9.陳雅雯 (2002)，「支援向量機於預測臺灣股市股價漲跌之實證研究」，南華大學資訊管理學研究所碩士論文。
10.黃俊霖 (2004)，「應用約略及理論與模糊理論於台股指數期貨漲跌幅之預測」，國立台灣科技大學資訊管理研究所碩士論文。
11.黃銘煌 (1999)，「TAIFEX 與SIMEX 台股指數期貨跨市場價差交易策略之研究」，國立台灣大學商學研究所碩士論文。
12.廖廣毅 (1999)，「以類神經網路預測股價指數漲跌」，私立元智大學工業工程研究所碩士論文。
13.賴佳君 (2006)，「應用平滑支撐向量機於台指期貨套利之研究」，國立台灣科技大學資訊管理研究所碩士論文。
英文部分
1.Adrangi, B., Chatrath, A., David, R.C. (2000), “Price Discovery in Strategically-linked Markets : The case of the gold-silver spread”, Applied Financial Economics, Vol.10, No.3, pp.227-234
2.Bergerson, K. and Wunsch, D. C. (1991),”A Commodity Trading Model Based on a Neural Network-expert Systermhybird”, Neural Trading, Vol.1, pp.289-293
3.Billingsley, R. S., and Chance, D. M. (1988), “The Pricing and Performance of Stock Index Futures Spreads”, Journal of Futures Markets ,Vol.8, pp.303-318
4.Board, J., and Sutcliffe, C. (1996), “The Dual Listing of Stock Index Futures : Arbitrage, Spread Arbitrage, and Currency Risk”, Journal of Futures Markets ,Vol.16, pp.29-54
5.Butterworth, D., and Holmes, P. (2002), “Intermarket Spread Trading : Evidence from UK Index Futures Market”, Applied Financial Economics ,Vol.12, pp.783-790
6.Cornell, B. and French, K. R. (1983), “The Pricing of Stock Index Futures”, Journal of Futures Markets, Vol.3, No.1, pp.1-14
7.Hsu, C. W, Chang, C. C., and Lin, C. J. (2003), “A Practical Guide to Support Vector Classification”, Available at http://www.csie.ntu.edu.tw/~cjlin/papers/guide/guide.pdf.
8.Huang, W., Nakamori, Y. and Wang, S.-Y. (2005), “Computers and Operations Research”, Vol.32, No.10, pp.2513-2522
9.Klemkosky, R.C. and Lee, J. H. (1991), “The Intraday Ex Post and Ex Ante Profitability of Index Arbitrage”, Journal of Futures Markets, Vol.11, No.3, pp.291-311
10.Kim, K. J. (2003), ”Financial Time Series Forecasting Using Support Vector Machines”, Neurocomputing, Vol.55, No.1-2 pp. 307-319
11.Koski, J. and Pontiff, J. (1996),”How Are Derivatives Used? Evidence from the Mutual Fund Industry”, Journal of Finance, Vol.54, No.2, pp.791-816
12.Lee, Y.-J., Mangasarian, O.L. (2001),”SSVM：A Smooth Support Vector Machine for Classification”, IEEE Transactions on Knowledge and Data Engineering, Vol.17, No.5, pp. 678-685.
13.Lee, Y.-J., Hsieh, W.-F., Huang, C.-M. (2005), ”ε-SSVR: A Smooth Support Vector Machine for ε-insensitive Regression”, IEEE Transactions on Knowledge and Data Engineering, Vol.17, No.5, pp. 678-685
14.Vapnik, V. N. (1995),”The Nature of Statistical Learning Theory”, Springer Verlan, New York.
15.Mary, T. P., and Tamvakis, M. N. (1999),”The dynamic relationship between paper petroleum refining and physical trade of crude oil into the Untied States”, Maritime Policy and Management, Vol.26, No.2, pp.127-136
16.Merrick, J. J. (1989),”Early Unwindings and Rollovers of Stock Index Futures Arbitrage Programs”, Journal of Futures Markets, Vol.9, pp.101-111
17.Modest, D.M. and Sundaresan, M. (1983), “The Relationship Between Spot and Futures Prices in Stock Index Futures Markets: Some Preliminary Evidence”, Journal of Futures Markets,Vol.3, No.1, pp.15-41
18.Sullivan, R. Timmermann, A. and White, H. (1999),”Data-Snooping, Technical Trading Rule Performance, and the Bootstrap”, Journal of Finance, Vol.54, No.5, pp.1647