頁籤選單縮合
題名 | Predicting Prepayment of Residential Mortgages: A Neural Network Approach |
---|---|
作者姓名(外文) | Waller,Bennie; Aiken,Milam; | 書刊名 | International Journal of Information and Management Sciences |
卷期 | 9:4 1998.12[民87.12] |
頁次 | 頁37-44 |
分類號 | 554.89 |
關鍵詞 | Neural network; Prepayment of mortgages; Logit model; |
語文 | 英文(English) |
英文摘要 | Prepayment of mortgages is one of the most important problems facing mortgage lending institutions as well as investors who purchase mortgages or mortgage-backed securities in secondary markets. While most researchers rely upon conventional statistical techniques to predict mortgage prepayments and mortgage lending institutions use arbitrary prepayment rules, neural networks may achieve greater prediction accuracies. Here, we examine the predictive validity of a neural network and compare it to a logit model. Using loan characteristic data, the neural network was able to predict prepayments of first residential mortgages with over 70% accuracy while the logit model achieved an accuracy rate of just over 50%. |
本系統之摘要資訊系依該期刊論文摘要之資訊為主。