国产日韩欧美一区二区三区三州_亚洲少妇熟女av_久久久久亚洲av国产精品_波多野结衣网站一区二区_亚洲欧美色片在线91_国产亚洲精品精品国产优播av_日本一区二区三区波多野结衣 _久久国产av不卡

?

基于灰色ARIMA的金融時(shí)間序列智能混合預(yù)測(cè)研究

2014-06-28 06:54羅洪奔
關(guān)鍵詞:灰色預(yù)測(cè)遺傳算法

羅洪奔

摘 要:提出了一種基于灰色ARIMA的金融時(shí)間序列智能混合預(yù)測(cè)模型。首先建立金融時(shí)間序列灰色預(yù)測(cè)模型,并采用PSO算法對(duì)灰色模型的三個(gè)參數(shù)進(jìn)行優(yōu)化;利用ARIMA算法對(duì)預(yù)測(cè)模型的殘差進(jìn)行分析,同時(shí)采用遺傳算法對(duì)ARIMA的系數(shù)進(jìn)行優(yōu)化;最后用ARIMA的殘差預(yù)測(cè)結(jié)果對(duì)灰色預(yù)測(cè)模型進(jìn)行補(bǔ)償。結(jié)果表明,以較好的精度擬合一段時(shí)期內(nèi)MA<107的時(shí)間序列,預(yù)測(cè)誤差控制在5%以上,與單純的灰色預(yù)測(cè)算法和神經(jīng)網(wǎng)絡(luò)算法相比,在平均絕對(duì)誤差、均方根誤差和趨勢(shì)準(zhǔn)確率三項(xiàng)評(píng)價(jià)指標(biāo)上,具有明顯優(yōu)勢(shì)。

關(guān)鍵詞: 金融時(shí)間序列;灰色預(yù)測(cè);ARIMA;PSO;遺傳算法

中圖分類號(hào):TP273+.23 文獻(xiàn)標(biāo)識(shí)碼: A 文章編號(hào):1003-7217(2014)02-0027-08

一、引 言

金融市場(chǎng)屬典型的復(fù)雜系統(tǒng),呈現(xiàn)出較強(qiáng)的非線性和時(shí)變性特征[1]。其內(nèi)部因素和位置變量之間的關(guān)系很難用準(zhǔn)確的數(shù)學(xué)公式加以描述, 難以建立完整的動(dòng)力方程。因此,研究針對(duì)金融時(shí)間序列的分析和預(yù)測(cè)方法,具有十分重要的意義。

近年來(lái),智能算法被越來(lái)越多地應(yīng)用于金融時(shí)間序列的預(yù)測(cè)中,如神經(jīng)網(wǎng)絡(luò)、支持向量機(jī)等算法。這類方法一定程度上能解決市場(chǎng)非線性、非平穩(wěn)性和高信噪比等問(wèn)題, 但由于訓(xùn)練速度慢,學(xué)習(xí)過(guò)程誤差容易陷入局部極小點(diǎn),很難保證學(xué)習(xí)精度。另外,這類方法只能保證在有限樣本的情況下經(jīng)驗(yàn)風(fēng)險(xiǎn)最小,而預(yù)測(cè)精度難以保證,泛化能力不高,應(yīng)用范圍受到了一定限制。

鑒此,本文從金融市場(chǎng)的特性和變化規(guī)律出發(fā),提出一種基于灰色-ARIMA的智能混合預(yù)測(cè)方法。將傳統(tǒng)方法與智能方法相結(jié)合,利用灰色理論建立金融時(shí)間序列模型。為了避免參數(shù)估計(jì)引入的誤差,采用粒子群算法(簡(jiǎn)稱PSO)對(duì)灰色模型參數(shù)進(jìn)行尋優(yōu),同時(shí)利用ARIMA算法對(duì)預(yù)測(cè)模型的殘差進(jìn)行分析,以達(dá)到消除殘差的目的;為進(jìn)一步提升算法的精度,采用遺傳算法對(duì)ARIMA的系數(shù)進(jìn)行估計(jì)。

二、金融時(shí)間序列預(yù)測(cè)算法的提出

金融時(shí)間序列呈現(xiàn)出的波動(dòng)性、非平穩(wěn)性、周期性、樣本少的特點(diǎn)對(duì)預(yù)測(cè)算法提出了較高的要求。鄧聚龍(1982)提出的灰色系統(tǒng)理論認(rèn)為,任何隨機(jī)過(guò)程都是在一定時(shí)區(qū)范圍內(nèi)變化的灰色過(guò)程,通過(guò)細(xì)分處理,可歸結(jié)為一種連續(xù)的、平穩(wěn)的、動(dòng)態(tài)的隨機(jī)過(guò)程[2]。對(duì)金融時(shí)間序列這類“貧信息,小樣本”問(wèn)題,灰色系統(tǒng)理論有較好的分析效果,能夠有效地分析金融時(shí)間序列變化地本質(zhì)規(guī)律和變化周期。

傳統(tǒng)的灰色預(yù)測(cè)方法在處理數(shù)據(jù)時(shí)都做了一定的條件假設(shè),這些假設(shè)在對(duì)于金融時(shí)間序列的預(yù)測(cè)問(wèn)題中不一定成立。同時(shí)金融活動(dòng)中存在大量由突發(fā)因素造成的波動(dòng)性和非平穩(wěn)性變化,這些變化可能無(wú)規(guī)律可循,灰色模型難以辨識(shí),使得辨識(shí)得到的金融時(shí)間序列模型與真實(shí)數(shù)據(jù)間存在較大的誤差,成為提高預(yù)測(cè)模型精度的瓶頸[3]。因此,本文提出一種混合預(yù)測(cè)模型(如圖1所示),該混合預(yù)測(cè)模型由灰色預(yù)測(cè)和殘差預(yù)測(cè)兩部分組成。改進(jìn)灰色預(yù)測(cè)模型,采用灰色理論的系統(tǒng)分析方法對(duì)原始金融時(shí)間序列數(shù)據(jù)進(jìn)行辨識(shí),從而逼近金融市場(chǎng)的變化規(guī)律。為了減少參數(shù)假設(shè)所引入的系統(tǒng)誤差,將灰色模型進(jìn)行擴(kuò)展,采用PSO方法對(duì)模型參數(shù)進(jìn)行優(yōu)化,提升模型精度。

圖1 預(yù)測(cè)算法原理圖

基于ARIMA算法的殘差預(yù)測(cè)模型,針對(duì)金融時(shí)間序列的波動(dòng)性和非平穩(wěn)性特點(diǎn),將灰色預(yù)測(cè)模型與原始序列間的殘差進(jìn)行分析,建立殘差預(yù)測(cè)模型,對(duì)灰色預(yù)測(cè)模型進(jìn)行修正。為降低ARIMA參數(shù)預(yù)估引入的預(yù)測(cè)風(fēng)險(xiǎn),引入遺傳算法進(jìn)行優(yōu)化。

三、灰色預(yù)測(cè)模型

金融時(shí)間序列的精確預(yù)測(cè)是保證金融市場(chǎng)高效、穩(wěn)定運(yùn)行的基礎(chǔ)。在灰色系統(tǒng)領(lǐng)域,金融活動(dòng)可以看作在一定時(shí)區(qū)、一定范圍內(nèi)變化的灰色過(guò)程。其本質(zhì)是:通過(guò)對(duì)歷史金融數(shù)據(jù)進(jìn)行累加生成,整理成規(guī)律性較強(qiáng)的數(shù)據(jù)序列,結(jié)合微分?jǐn)M合法建立微分方程來(lái)描述生成時(shí)間序列的規(guī)律,實(shí)現(xiàn)對(duì)將來(lái)時(shí)刻的預(yù)測(cè)。GM(1,1)是最為傳統(tǒng)的灰色預(yù)測(cè)算法,由于對(duì)于背景值選取做了一定假設(shè)和限制,造成預(yù)測(cè)誤差偏大。針對(duì)這一問(wèn)題,本文在GM(1,1)的基礎(chǔ)上引入向量β,將其推廣為GM(1,1,β)模型,并采用PSO算法優(yōu)化該模型的關(guān)鍵參數(shù),提高預(yù)測(cè)精度。

從表1的結(jié)果可以看出,對(duì)50個(gè)交易日的預(yù)測(cè)數(shù)據(jù)進(jìn)行分析,本文提出的組合預(yù)測(cè)算法在. RMSE、MAPE和F. 三項(xiàng)指標(biāo)遠(yuǎn)小于其他兩種預(yù)測(cè)算法,表明灰色ARIMA算法能夠?qū)W習(xí)和跟蹤股市變化情況,具有很好擬合能力。灰色ARIMA算法在50次趨勢(shì)預(yù)測(cè)中,正確趨勢(shì)45次,準(zhǔn)確率為90%,對(duì)未來(lái)具有很好的判斷能力,從而說(shuō)明灰色ARIMA算法對(duì)上證綜合指數(shù)的趨勢(shì)有很好的跟蹤能力。

六、結(jié)論

以上在對(duì)金融時(shí)間序列自身特點(diǎn)充分分析的基礎(chǔ)上,針對(duì)金融市場(chǎng)中存在的干擾因素眾多,關(guān)系復(fù)雜,呈現(xiàn)波動(dòng)性、非平穩(wěn)性,提出了一種灰色-ARIMA的金融時(shí)間序列的智能混合預(yù)測(cè)模型。

從實(shí)證分析的結(jié)果上看,本文算法能以較好的精度擬合一段時(shí)期內(nèi)金融時(shí)間序列數(shù)據(jù),由于采用了殘差消除和智能優(yōu)化方法,模型預(yù)測(cè)精度比單純的灰色預(yù)測(cè)算法有了較大提升,從而提供了一種新的分析金融時(shí)間序列的有效途徑。

參考文獻(xiàn):

[1]倪麗萍, 倪志偉. 一種基于趨勢(shì)分形維數(shù)的股指時(shí)間序列相似性分析方法[J]. 系統(tǒng)工程理論與實(shí)踐,2012,21(9):76-78.

[2]毛麗, 左青松, 劉冠麟. 車(chē)用三效催化轉(zhuǎn)化器剩余壽命的非等間隔灰色預(yù)測(cè)[J]. 中南大學(xué)學(xué)報(bào)(自然科學(xué)版), 2012, 16(4): 59-61.

[3]周?chē)?guó)雄, 吳 敏. 基于改進(jìn)的灰色預(yù)測(cè)的模糊神經(jīng)網(wǎng)絡(luò)預(yù)測(cè)[J]. 系統(tǒng)仿真學(xué)報(bào),2010,22(10):68-71.

[4]魏徳敏,文星宇. 基于混合PSO算法的桁架動(dòng)力響應(yīng)優(yōu)化[J]. 振動(dòng)與沖擊, 2011, 22(5): 92-95.

[5]黃安強(qiáng), 肖進(jìn), 汪壽陽(yáng). 一個(gè)基于集成情境知識(shí)的組合預(yù)測(cè)方法[J]. 系統(tǒng)工程理論與實(shí)踐, 2011, (1):123-127.

[6]A. Azaron, C. Perkgoz, M. Sakawa. A genetic algorithm approach for the timecost tradeoff in PERT networks[J]. Applied Mathematics and Computation, 2005, 168 (2): 1317-1339.

[7]李松, 劉力軍, 解永樂(lè). 遺傳算法優(yōu)化BP神經(jīng)網(wǎng)絡(luò)的短時(shí)交通流混沌預(yù)測(cè)[J]. 控制與決策, 2011, 26(5): 76-81.

(責(zé)任編輯:姚德權(quán))

An Intelligent Hybrid Prediction for Financial Time Series Based on the GreyARIMA

LUO Hongben1,2

. (1.School of Business,Central South University, Changsha, Hunan 410083,China;

2.Office of Scientific R&D Hunan University,Changsha,Hunan 410082,China).

Abstract:An Intelligent hybrid financial time series forecasting model is proposed based on a grey ARIMA. First, the financial times series grey forecasting model is constructed, and at the same time three parameters were optimized using PSO algorithm. The grey forecasting model residuals are then analyzed with ARIMA, and the coefficients for the ARIMA model are optimized with a genetic algorithm. Finally, the predicative results of the ARIMA model are used to compensate the grey forcasting model.The empirical results show that the algorithm proposed in this paper can have better fitting precision for a period of MA<107 time series data with the prediction error controlled within 5%; compared with the grey prediction algorithm and the neural network algorithm, the algorithm has obvious advantages in terms of the mean absolute error, root mean square error and the trend prediction.

Key words:Financial Time Series; Grey Prediction; ARIMA; PSO; Genetic Algorithm

[5]黃安強(qiáng), 肖進(jìn), 汪壽陽(yáng). 一個(gè)基于集成情境知識(shí)的組合預(yù)測(cè)方法[J]. 系統(tǒng)工程理論與實(shí)踐, 2011, (1):123-127.

[6]A. Azaron, C. Perkgoz, M. Sakawa. A genetic algorithm approach for the timecost tradeoff in PERT networks[J]. Applied Mathematics and Computation, 2005, 168 (2): 1317-1339.

[7]李松, 劉力軍, 解永樂(lè). 遺傳算法優(yōu)化BP神經(jīng)網(wǎng)絡(luò)的短時(shí)交通流混沌預(yù)測(cè)[J]. 控制與決策, 2011, 26(5): 76-81.

(責(zé)任編輯:姚德權(quán))

An Intelligent Hybrid Prediction for Financial Time Series Based on the GreyARIMA

LUO Hongben1,2

. (1.School of Business,Central South University, Changsha, Hunan 410083,China;

2.Office of Scientific R&D Hunan University,Changsha,Hunan 410082,China).

Abstract:An Intelligent hybrid financial time series forecasting model is proposed based on a grey ARIMA. First, the financial times series grey forecasting model is constructed, and at the same time three parameters were optimized using PSO algorithm. The grey forecasting model residuals are then analyzed with ARIMA, and the coefficients for the ARIMA model are optimized with a genetic algorithm. Finally, the predicative results of the ARIMA model are used to compensate the grey forcasting model.The empirical results show that the algorithm proposed in this paper can have better fitting precision for a period of MA<107 time series data with the prediction error controlled within 5%; compared with the grey prediction algorithm and the neural network algorithm, the algorithm has obvious advantages in terms of the mean absolute error, root mean square error and the trend prediction.

Key words:Financial Time Series; Grey Prediction; ARIMA; PSO; Genetic Algorithm

[5]黃安強(qiáng), 肖進(jìn), 汪壽陽(yáng). 一個(gè)基于集成情境知識(shí)的組合預(yù)測(cè)方法[J]. 系統(tǒng)工程理論與實(shí)踐, 2011, (1):123-127.

[6]A. Azaron, C. Perkgoz, M. Sakawa. A genetic algorithm approach for the timecost tradeoff in PERT networks[J]. Applied Mathematics and Computation, 2005, 168 (2): 1317-1339.

[7]李松, 劉力軍, 解永樂(lè). 遺傳算法優(yōu)化BP神經(jīng)網(wǎng)絡(luò)的短時(shí)交通流混沌預(yù)測(cè)[J]. 控制與決策, 2011, 26(5): 76-81.

(責(zé)任編輯:姚德權(quán))

An Intelligent Hybrid Prediction for Financial Time Series Based on the GreyARIMA

LUO Hongben1,2

. (1.School of Business,Central South University, Changsha, Hunan 410083,China;

2.Office of Scientific R&D Hunan University,Changsha,Hunan 410082,China).

Abstract:An Intelligent hybrid financial time series forecasting model is proposed based on a grey ARIMA. First, the financial times series grey forecasting model is constructed, and at the same time three parameters were optimized using PSO algorithm. The grey forecasting model residuals are then analyzed with ARIMA, and the coefficients for the ARIMA model are optimized with a genetic algorithm. Finally, the predicative results of the ARIMA model are used to compensate the grey forcasting model.The empirical results show that the algorithm proposed in this paper can have better fitting precision for a period of MA<107 time series data with the prediction error controlled within 5%; compared with the grey prediction algorithm and the neural network algorithm, the algorithm has obvious advantages in terms of the mean absolute error, root mean square error and the trend prediction.

Key words:Financial Time Series; Grey Prediction; ARIMA; PSO; Genetic Algorithm

猜你喜歡
灰色預(yù)測(cè)遺傳算法
遺傳算法對(duì)CMAC與PID并行勵(lì)磁控制的優(yōu)化
基于自適應(yīng)遺傳算法的CSAMT一維反演
一種基于遺傳算法的聚類分析方法在DNA序列比較中的應(yīng)用
基于遺傳算法和LS-SVM的財(cái)務(wù)危機(jī)預(yù)測(cè)
2016年世界園藝博覽會(huì)對(duì)唐山經(jīng)濟(jì)的影響
收益還原法在房地產(chǎn)估價(jià)工作中的應(yīng)用與改進(jìn)
協(xié)同進(jìn)化在遺傳算法中的應(yīng)用研究
全面放開(kāi)二胎政策對(duì)蚌埠市人口結(jié)構(gòu)的影響的探索
基于灰色預(yù)測(cè)模型的中國(guó)汽車(chē)市場(chǎng)預(yù)測(cè)
組合預(yù)測(cè)法在我國(guó)糧食產(chǎn)量預(yù)測(cè)中的應(yīng)用