本研究使用機器學習模型預測擔保房貸憑證(Collateral Mortgage Obligation, CMOs)之提前還款率並評價，且和两种傳统的提前還款率的模型進行比較。第一種是靜態的提前還款模型，使用聯邦住宅管理局經驗法(Federal Home Administration, FHA)、條件提前還款率(Conditional Prepayment Rate, CPR)或以美國公共證券協會(The Public Securities Association, PSA)提前還款基準作為提前還款預測的模型。第二種是動態的提前還款，由美國儲蓄機構管理局(Office Thrift Supervision, OTS)提出的30年期固定利率房屋抵押貸款動態提前還款模型。由於評價CMOs時會將現金流進行折現，且票面利息的計算會使用到倫敦銀行同業隔夜拆款利率(London Interbank Offered Rate, Libor)。因此，本研究使用Hull & White利率模型模擬即期利率路徑，再通過遠期利率協定(Forward Rate Agreement, FRA)轉換成遠期Libor的路徑計算現金流。通過Fannie Mae發行的一檔CMOs商品的公開資料用於實證，實證結果證實機器學習預測提前還款優於傳统模型。

In this paper we predict the prepayment rate and price the Collateral Mortgage Obligation by using Machine Learning, and compare the results with two traditional prepayment models. The first one is static prepayment model, which uses Federal Housing Administration (FHA) Model, Conditional Prepayment Rate (CPR) Model or the Public Securities Association (PSA) prepayment benchmark for the prepayment model. The second one is the dynamic prepayment model from Office Thrift Supervision (OTS), which uses 30 years fixed mortgage rate. Because the high relationship between coupon rate of CMO trench and Libor rate, this paper uses Hull & White interest rate model to simulate the spot interest rate as the discount rate, and converts it to the Libor rate with the help of Forward Rate Agreement (FRA). The empirical analysis based on a CMOs issued by Fannie Mae illustrated that for Machine Learning, the efficiency in predicting the prepayment rate is better than traditional models.

中文文獻
1. 王立偉，2008，「提前還款對住房抵押貸款支持證券定價影響的效果」，大連理工大學金融工程系碩士班碩士論文。
2. 高心怡，2000，「結合HULL-WHITE利率模型與PHM提前清償模型評價CMO利率衍生性商品」，國立台灣大學財務金融系碩士班碩士論文。
3. 張繼文，2010，「擔保房貸憑證（CMOs）評價-以BGM利率模型為例」，國立政治大學金融系碩士班碩士論文。
4. 張憲明，2018，「擔保房貸憑證（CMOs）之評價：應用類神經網路預測提前還款率」，國立政治大學金融系碩士班碩士論文。
5. 廖伯媛，2001，「不動產抵押貸款證券化之分析與評價」，國立政治大學金融系碩士班碩士論文。
6. 劉展宏、張金鶚，2001，「購屋貸款提前清償行為之研究」，住宅學報，10 卷 1 期：29~49。

英文文獻
1. Andreas, K, and Rudi, Z, 2008,”A Hybrid-Form Model for the Prepayment-Risk-Neutral Valuation of Mortgage-Backed Securities”, The International Journal of Theoretical and Applied Finance, 11, pp.635-656
2. Clauretie and Sirmans, 1999, “Real Estate Finance Theory Practice”, Longman Higher Education Division ; 3rd edition
3. Deng Y., 1977, ”Mortgage Termination: An Empirical Hazard Model with Stochastic Term Structure ”, Journal of Real Estate Finance and Economics,Vol.14,pp.309-331
4. Dunn, K. B. and McConnell, J. J. (1981), “Valuation of GINNIE MAE Mortgage-Backed Securities,” Journal of Finance, Vol.36, pp.599-616
5. Dunn, K. B. and McConnell, J. J. (1981), “A Compare of Alternative Models for Pricing GINNIE MAE Mortgage-Back Securities,” Journal of Finance Vol.36, pp.471-484.
6. Green ,J , and J. B. Shoven ,1983,”The Effect of Interest Rates on Mortgage Prepayments”, Journal of Money, Credit and Banking 18,41-59
7. Gurrieri , M. Nakabayashi & T. Wong (2009), “Calibration methods of Hull–White model”, Working paper.
8. Jone, J. Mcconnell and Manoj Singh,1993,”Valuation and Analysis of Collateralized Mortgage Obligations”, Management Science, Vol.39, No. 6,pp.692-709
9. Jone J. Mcconnell and Manoj Singh, 1994,”Rational Prepayments and the Valuation of Collateralized Mortgage Obligations”, The Journal of Finance, Vol. 49, No. 3,pp.891-921
10. Ronald w. Spahr and Mark A. Sunderman,”The Effect of Prepayment Modeling in Pricing Mortgage-Backed Securities”, Journal of Housing Research, Vol. 3,pp.381-400
11. R Riksen , 2017,”Using Articial Neural Networks in the Calculation of Mortgage Prepayment Risk”, University of Amsterdam, Korteweg-de Vries Institute for Mathematics
12. Scott F. Richard Roll, 1989,”Prepayments on Fixed-Rate Mortgage-Backed Securities”, Journal of Portfolio Management 15,pp.73-82
13. Schwartz, Eduardo S., and Walter N. Torous, 1989, "Prepayment and the Valuation of Mortgage-Backed Securities," Journal of Finance, 44, 375-39
14. Waller B. and Aiken M.,1998, “Predicting Prepayment of Residential Mortgages: A Neural Network Approach ”, Information and Management Sciences, 9, pp.37-44