【摘要】：混沌是自相對(duì)論和量子力學(xué)之后人類科學(xué)的又一偉大發(fā)現(xiàn),是對(duì)人類整個(gè)知識(shí)體系的又一次巨大沖擊。它改變了自牛頓體系確定以來人們對(duì)于整個(gè)知識(shí)體系的認(rèn)知,擴(kuò)展了人們對(duì)于自然界的認(rèn)識(shí)�；煦缡且环N低階確定性的非線性動(dòng)力系統(tǒng)所表現(xiàn)出來的非常復(fù)雜的行為,動(dòng)力系統(tǒng)長(zhǎng)期演化中任一變量的演化過程都包含了系統(tǒng)所有變量的信息,通過單變量時(shí)間序列反向構(gòu)造出原系統(tǒng)相空間結(jié)構(gòu)。 中國(guó)股票市場(chǎng)自20世紀(jì)90年代成立,經(jīng)過20多年的發(fā)展,曾經(jīng)歷了幾次暴漲暴跌。國(guó)內(nèi)外研究表明,股票市場(chǎng)存在混沌現(xiàn)象,而中國(guó)股票市場(chǎng)是一個(gè)具有分形維結(jié)構(gòu)的混沌系統(tǒng)。而與股票市場(chǎng)有著重要關(guān)系的國(guó)際原油市場(chǎng)也被研究證明了是一個(gè)存在混沌現(xiàn)象的復(fù)雜系統(tǒng)。本文在對(duì)股票市場(chǎng)研究預(yù)測(cè)模型進(jìn)行了簡(jiǎn)單概述后,使用BP神經(jīng)網(wǎng)絡(luò)訓(xùn)練預(yù)測(cè)模型對(duì)中國(guó)股票市場(chǎng)上證指數(shù)收盤價(jià)進(jìn)行訓(xùn)練預(yù)測(cè),同時(shí)對(duì)國(guó)際原油期貨價(jià)格市場(chǎng)WTI即美國(guó)西德克薩斯輕質(zhì)原油價(jià)格使用了同樣的BP神經(jīng)網(wǎng)絡(luò)模型進(jìn)行了訓(xùn)練預(yù)測(cè)分析。 本文對(duì)混沌現(xiàn)象發(fā)現(xiàn)過程進(jìn)行了梳理,從圖靈模式到別洛烏索夫反應(yīng),從《確定性的非周期流》到邏輯斯蒂方程,從費(fèi)根鮑姆常數(shù)到曼德勃羅集,從自組織到自相似。對(duì)混沌學(xué)理論基礎(chǔ)做了理論分析。之后簡(jiǎn)述了中國(guó)股票市場(chǎng)發(fā)展歷史,對(duì)股票指數(shù)的計(jì)算方法和對(duì)中國(guó)股票市場(chǎng)預(yù)測(cè)研究現(xiàn)狀做了簡(jiǎn)述。最后使用混沌理論方法對(duì)上證指數(shù)收盤價(jià)和WTI原油價(jià)格進(jìn)行混沌分析,互信息函數(shù)方法求取延遲時(shí)間,CAO方法求取嵌入維數(shù),對(duì)其進(jìn)行了相空間重構(gòu)。并對(duì)上證指數(shù)收盤價(jià)根據(jù)G-P算法對(duì)上證指數(shù)收盤價(jià)計(jì)算其關(guān)聯(lián)維為非整數(shù),使用最小數(shù)據(jù)量方法計(jì)算對(duì)這兩個(gè)經(jīng)濟(jì)數(shù)據(jù)計(jì)算最大Lyapunov指數(shù)都大于零,由此可以得出我國(guó)股票市場(chǎng)和國(guó)際原油市場(chǎng)是一個(gè)存在混沌現(xiàn)象的復(fù)雜系統(tǒng),并對(duì)其建立了基于最大Lyapunov指數(shù)混沌預(yù)測(cè)模型。通過和BP神經(jīng)網(wǎng)絡(luò)訓(xùn)練模型進(jìn)行對(duì)比分析,發(fā)現(xiàn)基于最大Lyapunov指數(shù)混沌預(yù)測(cè)模型具有較好的預(yù)測(cè)效果。
[Abstract]:Chaos is another great discovery of human science since relativity and quantum mechanics, and it is another great impact on the whole human knowledge system. It has changed people's cognition of the whole knowledge system since Newton's system was determined, and expanded people's understanding of nature. Chaos is a very complex behavior of a low-order deterministic nonlinear dynamic system. The evolution process of any variable in the long-term evolution of the dynamical system contains the information of all the variables of the system. The phase space structure of the original system is inversely constructed by univariate time series. After more than 20 years of development, China's stock market has experienced several spikes and plunges since its establishment in the 1990 s. Studies at home and abroad show that there is chaos in the stock market, while the Chinese stock market is a chaotic system with fractal dimension structure. The international crude oil market, which has an important relationship with the stock market, has also been proved to be a complex system with chaotic phenomena. In this paper, after a brief overview of the stock market research forecasting model, the BP neural network training forecasting model is used to forecast the closing price of the Shanghai Stock Exchange Index in China stock market. At the same time, the same BP neural network model is used to predict the international crude oil futures market WTI, that is, the West Texas light crude oil price. In this paper, the discovery process of chaotic phenomena from Turing model to Belousov reaction, from deterministic aperiodic flow to logical Stey equation, from Fegenbaum constant to Manderborough set, from self-organization to self-similarity is discussed. The theoretical basis of chaos is analyzed. Then, the history of Chinese stock market development, the calculation method of stock index and the present situation of stock market prediction in China are briefly described. Finally, chaos theory is used to analyze the closing price of Shanghai Stock Exchange Index and the price of WTI crude oil, the delay time is obtained by mutual information function method, and the embedding dimension is obtained by CAO method, and the phase space is reconstructed. On the basis of G-P algorithm, the correlation dimension of the closing price of Shanghai Stock Exchange Index is calculated as a non-integer, and the maximum Lyapunov index of the two economic data is calculated by the method of minimum data. It is concluded that the stock market in China and the international crude oil market are a complex system with chaotic phenomena, and a chaotic prediction model based on the largest Lyapunov exponent is established. By comparing with the BP neural network training model, it is found that the chaotic prediction model based on the maximum Lyapunov exponent has better prediction effect.
【學(xué)位授予單位】：東北林業(yè)大學(xué)
【學(xué)位級(jí)別】：碩士
【學(xué)位授予年份】：2012
【分類號(hào)】：F224;F832.51

【引證文獻(xiàn)】

相關(guān)碩士學(xué)位論文 前1條

1 詹財(cái)鑫;基于SVM_AdaBoost模型的股票漲跌實(shí)證研究[D];華南理工大學(xué);2013年

，

本文編號(hào)：2362848