頁籤選單縮合
題名 | 計算智慧於臺股認購權證定價的可行性評估=The Evaluation of Computational Intelligence in Taiwan Stock Call Warrant Pricing |
---|---|
作者 | 李沃牆; | 書刊名 | 證券市場發展季刊 |
卷期 | 9:4=36 1997[民86.] |
頁次 | 頁89-116 |
分類號 | 563.538 |
關鍵詞 | 認購權證; 選擇權; 訂價模型; 二項式訂價模型; 遺傳演算法; 人工神經網路; Call warrant; Option; Black-scholes option pricing; Genetic apgorithms; Artificial neural network; |
語文 | 中文(Chinese) |
中文摘要 | 國內首批認購權證已於86年9月初掛牌上市,截至目前為止,已有12檔權證上市掛牌交易,認購權證在國外早已不算新鮮,但在國內卻屬一項新的金融衍生性商品,此產品不但為國內企業,更為個人投資者帶來新的避險及投資工具。 認購權證在設計上是一種選擇權,故其定價適用選擇權的定價模式,傳統的定價方法不外是以Black-Scholes的定價模型或是二項式評價模型為主再加上其它的數值演算法,但這些模型在實務應用上都有其限制,本文首先應用遺傳演算法做為數值分析解,其逼近效果相當不錯。另外,在實證部分,則用無母數的人工神經網路,在考慮成交量,不同的歷史波動性及隱藏層單元數的情況下,進行臺股認購權證的評價,實證結果顯示,神經網路確實能用於認購權證評價,並獲致不錯的績效。 |
英文摘要 | The first stock call warrant has been trade on Taiwan security market on September, 1997. Till now, it has fourteen security broker with the right to development. Call warrant is a new derivative security in domestic. It bring the enterprise, agent a new investment tool. Call warrant is a option in design. That is to say, the option pricing is fittable in call warrant. Traditional methods include Balck-Scholes option pricing model, binomial model and other numerical methods. But they face some limits in empirical study. So, the purpose of my paper, try to use gentic algorithms as a numerical solution, potential to deal with warrant pricing. On the other side, I also use the artificial neural network as a nonparametric method, consider the trade volume, differential historical volatility and hidden units to evaluate the call warrantprice. Results show that it can get well performance. |
本系統之摘要資訊系依該期刊論文摘要之資訊為主。