發(fā)布時間：2018-07-23 15:15

【摘要】：本文在對神經(jīng)網(wǎng)絡(luò)模型與VAR方法的應(yīng)用及實(shí)現(xiàn)進(jìn)行詳細(xì)介紹的基礎(chǔ)上,討論了應(yīng)用神經(jīng)網(wǎng)絡(luò)模型與VAR對中國銀行匯率風(fēng)險管理進(jìn)行優(yōu)化的可行性,形成一套具有一定實(shí)際意義的風(fēng)險管理優(yōu)化方案。 本文首先對文中所采用的兩個風(fēng)險管理模型進(jìn)行了介紹。在對神經(jīng)網(wǎng)絡(luò)模型進(jìn)行介紹時,選取了2009年12月31日至2013年3月1日共760個人民幣對美元中間價作為樣本,利用MATLAB軟件建立神經(jīng)網(wǎng)絡(luò)模型,通過建立模型、訓(xùn)練模型、測試模型等三個步驟實(shí)現(xiàn)神經(jīng)網(wǎng)絡(luò)模型對外匯中間價的預(yù)測,獲得的測試結(jié)果與實(shí)際數(shù)據(jù)基本吻合,證明了神經(jīng)網(wǎng)絡(luò)模型在匯率預(yù)測方面具有較高的準(zhǔn)確度。在對VAR方法進(jìn)行介紹時,選取2009年12月31日至2013年1月4日共724個交易日人民幣對美元外匯中間價作為實(shí)證樣本,采用報酬率計(jì)算公式為預(yù)測公式,獲得資產(chǎn)組合價值分布圖,最后以99%的置信區(qū)間獲得VAR數(shù)值,經(jīng)過以上四個步驟實(shí)現(xiàn)VAR的測算。 在完成對神經(jīng)網(wǎng)絡(luò)模型與VAR實(shí)現(xiàn)的介紹基礎(chǔ)上,從風(fēng)險管理組織架構(gòu)、外匯交易市場風(fēng)險管理、外匯交易操作風(fēng)險管理等三個角度,對中國銀行現(xiàn)行的外匯風(fēng)險管理系統(tǒng)進(jìn)行詳細(xì)分析,并利用神經(jīng)網(wǎng)絡(luò)模型與VAR從以上三個角度對中國銀行風(fēng)險管理進(jìn)行適用性分析,得出神經(jīng)網(wǎng)絡(luò)模型與VAR對中國銀行風(fēng)險管理的優(yōu)化具有適用性的結(jié)論。在以上評價的基礎(chǔ)上,將神經(jīng)網(wǎng)絡(luò)模型與VAR與中國銀行風(fēng)險管理現(xiàn)狀相結(jié)合,從中國銀行風(fēng)險管理組織架構(gòu)、市場風(fēng)險管理、操作風(fēng)險管理三個方面進(jìn)行優(yōu)化,與原有風(fēng)險管理管理系統(tǒng)進(jìn)行對比,優(yōu)化后的風(fēng)險管理系統(tǒng)有助于提高中國銀行的外匯風(fēng)險管理水平,具有良好的應(yīng)用潛能。
[Abstract]:Based on the detailed introduction of the application and implementation of neural network model and VAR method, this paper discusses the feasibility of optimizing the exchange rate risk management of Bank of China by using neural network model and VAR. To form a set of practical significance of risk management optimization scheme. Firstly, two risk management models are introduced in this paper. In the course of introducing the neural network model, 760 RMB / US dollar median prices from December 31, 2009 to March 1, 2013 are selected as samples. The neural network model is established by using MATLAB software, and the training model is established through the establishment of the neural network model. The neural network model is used to predict the intermediate value of foreign exchange rate. The test results are in good agreement with the actual data, which proves that the neural network model has a high accuracy in the prediction of exchange rate. When introducing the VAR method, 724 trading days from December 31, 2009 to January 4, 2013, are selected as the empirical samples, and the return formula is used as the forecast formula to obtain the portfolio value distribution map. Finally, the VAR value is obtained by 99% confidence interval, and the VAR is calculated through the above four steps. On the basis of the introduction of neural network model and VAR implementation, from three angles of risk management organization structure, foreign exchange trading market risk management, foreign exchange trading operation risk management, The current foreign exchange risk management system of Bank of China is analyzed in detail, and the applicability of risk management of Bank of China is analyzed from the above three angles by using neural network model and VAR. It is concluded that the neural network model and VAR are applicable to the optimization of Chinese bank risk management. On the basis of the above evaluation, combining the neural network model with VAR and the current situation of risk management of Bank of China, the paper optimizes the risk management of Bank of China from three aspects: organizational structure, market risk management and operational risk management. Compared with the original risk management system, the optimized risk management system is helpful to improve the level of foreign exchange risk management of Bank of China, and has good application potential.
【學(xué)位授予單位】：北京交通大學(xué)
【學(xué)位級別】：碩士
【學(xué)位授予年份】：2013
【分類號】：TP183;F832.33

【參考文獻(xiàn)】

相關(guān)期刊論文 前8條

1 楊p，

本文編號：2139803