發(fā)布時(shí)間：2018-09-07 12:19

【摘要】：匯率預(yù)測(cè)是一個(gè)重要的經(jīng)濟(jì)問(wèn)題,已經(jīng)引起了廣泛的關(guān)注。然而,外匯市場(chǎng)是一個(gè)多變量的非線性系統(tǒng),并且外匯市場(chǎng)中的各因素間的相關(guān)性錯(cuò)綜復(fù)雜。因此,匯率預(yù)測(cè)是一項(xiàng)重要而富有挑戰(zhàn)性的研究。神經(jīng)網(wǎng)絡(luò)作為非線性動(dòng)力學(xué)系統(tǒng),具有廣泛的適應(yīng)能力,學(xué)習(xí)能力,被成功地用于多變量非線性系統(tǒng)的建模和控制。 20世紀(jì)90年代以來(lái),神經(jīng)網(wǎng)絡(luò)在經(jīng)濟(jì)、金融領(lǐng)域得到廣泛的應(yīng)用,已經(jīng)成為匯率預(yù)測(cè)領(lǐng)域的有效工具之一。前饋神經(jīng)網(wǎng)絡(luò)(FFNN)是一種常用的匯率預(yù)測(cè)算法,但它的缺點(diǎn)是學(xué)習(xí)過(guò)程中易于陷入局部極小。深度信度網(wǎng)絡(luò)(DBN)是2006年新提出的一種神經(jīng)網(wǎng)絡(luò),能夠收斂到全局最優(yōu),從而得到更精確的預(yù)測(cè)結(jié)果。 本文綜述了匯率預(yù)測(cè)和深度信度網(wǎng)絡(luò)的理論框架,研究了DBN的學(xué)習(xí)算法,并通過(guò)實(shí)驗(yàn)設(shè)計(jì)出DBN的最優(yōu)網(wǎng)絡(luò)結(jié)構(gòu)。在此基礎(chǔ)上,首次提出基于DBN的匯率預(yù)測(cè)方法,進(jìn)行了相關(guān)實(shí)驗(yàn),并對(duì)實(shí)驗(yàn)結(jié)果進(jìn)行了分析。首先,對(duì)三種匯率序列數(shù)據(jù)做預(yù)處理,在訓(xùn)練階段,我們將深度信度網(wǎng)絡(luò)(DBN)與共軛梯度算法相結(jié)合,加快學(xué)習(xí)速度。測(cè)試階段,使用四種評(píng)價(jià)指標(biāo)來(lái)衡量算法的預(yù)測(cè)效果。最后將預(yù)測(cè)結(jié)果與前饋神經(jīng)網(wǎng)絡(luò)等幾種經(jīng)典算法的結(jié)果對(duì)比。實(shí)驗(yàn)結(jié)果表明,將DBN與共軛梯度法結(jié)合后,匯率預(yù)測(cè)的效果最好,具有良好的發(fā)展前景。
[Abstract]:Exchange rate forecasting is an important economic issue, which has attracted wide attention. However, the foreign exchange market is a multivariable nonlinear system, and the correlation between various factors in the foreign exchange market is complicated. Therefore, exchange rate forecasting is an important and challenging study. As a nonlinear dynamical system, neural network has wide adaptability and learning ability, and has been successfully used in modeling and control of multivariable nonlinear systems. The financial field has been widely used and has become one of the effective tools in the field of exchange rate forecasting. Feedforward neural network (FFNN) is a common exchange rate prediction algorithm, but its disadvantage is that it is easy to fall into local minima in the learning process. The deep reliability network (DBN) is a new neural network proposed in 2006, which can converge to the global optimum and obtain more accurate prediction results. In this paper, the theoretical framework of exchange rate prediction and depth reliability network is reviewed, the learning algorithm of DBN is studied, and the optimal network structure of DBN is designed through experiments. On this basis, the method of exchange rate forecasting based on DBN is put forward for the first time, and relevant experiments are carried out, and the experimental results are analyzed. Firstly, we preprocess the three kinds of exchange rate sequence data. In the training stage, we combine (DBN) with conjugate gradient algorithm to accelerate the learning speed. In the testing stage, four evaluation indexes are used to measure the prediction effect of the algorithm. Finally, the prediction results are compared with the results of several classical algorithms such as feedforward neural networks. The experimental results show that the combination of DBN and conjugate gradient method has the best effect and has a good prospect.
【學(xué)位授予單位】：南京大學(xué)
【學(xué)位級(jí)別】：碩士
【學(xué)位授予年份】：2012
【分類(lèi)號(hào)】：TP18;F830.7

【參考文獻(xiàn)】

相關(guān)期刊論文 前3條

1 李麗;;基于ARMA-GARCH模型的股市量?jī)r(jià)動(dòng)態(tài)關(guān)系研究[J];統(tǒng)計(jì)與決策;2011年04期

2 嚴(yán)太山;;基于遺傳算法的神經(jīng)網(wǎng)絡(luò)學(xué)習(xí)算法研究[J];湖南理工學(xué)院學(xué)報(bào)(自然科學(xué)版);2007年01期

3 黃丹,黃伯元;人工神經(jīng)網(wǎng)絡(luò)及其在醫(yī)療衛(wèi)生中的應(yīng)用[J];衛(wèi)生職業(yè)教育;2005年04期

，

本文編號(hào)：2228247