本文選題：時(shí)間序列預(yù)測(cè)　切入點(diǎn)：多任務(wù)學(xué)習(xí)　出處：《太原理工大學(xué)》2017年碩士論文

【摘要】：時(shí)間序列是一種常見的數(shù)據(jù)結(jié)構(gòu),廣泛地存在于人類社會(huì)活動(dòng)和客觀世界中。時(shí)間序列預(yù)測(cè)旨在依據(jù)時(shí)間序列中隱含的內(nèi)在變化規(guī)律建立數(shù)學(xué)模型,挖掘其隱含的時(shí)間演變關(guān)系,從而實(shí)現(xiàn)對(duì)序列未來(lái)發(fā)展趨勢(shì)的預(yù)測(cè)。深入分析來(lái)源于不同觀測(cè)事物的時(shí)間序列數(shù)據(jù),挖掘蘊(yùn)含的可用信息,對(duì)辨識(shí)事物發(fā)展趨勢(shì)、刻畫事物間的相關(guān)特性以及對(duì)事物變化的決策與把控等有重要的實(shí)際意義。目前,基于數(shù)據(jù)驅(qū)動(dòng)建模的時(shí)間序列預(yù)測(cè)方法可分為全局模型和局部模型。由現(xiàn)實(shí)世界得到的時(shí)間序列數(shù)據(jù)往往具有強(qiáng)非線性和不確定性等特點(diǎn),這就使得傳統(tǒng)的預(yù)測(cè)模型難以有效建模,從而限制了模型的預(yù)測(cè)精度�；谌繗v史數(shù)據(jù)建模的全局模型雖然方法簡(jiǎn)單,但是對(duì)于序列中的異常值有較高的敏感性。針對(duì)上述問(wèn)題,本文提出了一種基于云模型相似性度量的局部建模方法,并結(jié)合BP神經(jīng)網(wǎng)絡(luò)(BPNN)和最小二乘支持向量機(jī)(LS-SVM)建立預(yù)測(cè)模型,可有效提高序列的預(yù)測(cè)精度。上述局部建模方法雖然對(duì)預(yù)測(cè)精度有一定的貢獻(xiàn),但是仍屬于單任務(wù)學(xué)習(xí)的范疇,對(duì)時(shí)間序列中隱含的相關(guān)性信息沒(méi)有進(jìn)行充分挖掘,從而影響了模型的泛化性能。而多任務(wù)學(xué)習(xí)(MTL)能兼顧任務(wù)間的相關(guān)性與差異性,挖掘了任務(wù)間具有共享知識(shí)結(jié)構(gòu)的共性信息,最終改善了所有任務(wù)的學(xué)習(xí)性能。因次,本文綜合考慮局部模型與多任務(wù)學(xué)習(xí)的優(yōu)勢(shì),提出了一種基于多任務(wù)學(xué)習(xí)的局部建模方法。該方法將時(shí)間序列的相鄰時(shí)間點(diǎn)中蘊(yùn)含的多種信息可看成是多任務(wù)學(xué)習(xí)中的不同任務(wù)同時(shí)進(jìn)行學(xué)習(xí),從而提高模型的泛化能力。本文的研究工作主要包括以下幾個(gè)方面:(1)分析了幾種常見時(shí)間序列相似性度量方法的優(yōu)勢(shì)和不足,并對(duì)兩種不同類型的時(shí)間序列預(yù)測(cè)方法進(jìn)行了對(duì)比;(2)針對(duì)原始時(shí)間序列數(shù)據(jù)具有非線性、復(fù)雜性以及不確定性等問(wèn)題,采用云模型理論對(duì)原始序列以及一階差分處理后的序列同時(shí)進(jìn)行時(shí)間序列表示;(3)針對(duì)傳統(tǒng)距離函數(shù)難以有效度量包含不確定性因素的數(shù)據(jù),提出了基于云模型相似性度量的局部建模方法,并采用BPNN和LS-SVM建立預(yù)測(cè)模型進(jìn)行建模和預(yù)測(cè);(4)針對(duì)單任務(wù)學(xué)習(xí)方法在處理時(shí)間序列中存在的信息挖掘不充分、預(yù)測(cè)精度低等問(wèn)題,在提出的局部建模方法的框架下,將多任務(wù)學(xué)習(xí)用于時(shí)間序列預(yù)測(cè),提出了基于多任務(wù)學(xué)習(xí)的局部建模方法;(5)采用選自阿爾托大學(xué)理學(xué)院的機(jī)器學(xué)習(xí)應(yīng)用研究組的六個(gè)真實(shí)的工程數(shù)據(jù)集對(duì)提出的方法進(jìn)行了驗(yàn)證。
[Abstract]:Time series is a common data structure, which widely exists in human social activities and objective world.The purpose of time series prediction is to establish a mathematical model based on the inherent law of change implied in time series, and to excavate its implicit time evolution relationship, so as to predict the trend of future development of time series.It is of great practical significance to analyze the time series data from different observational objects and to mine the available information to identify the development trend of things, to depict the related characteristics of things, and to make and control the changes of things.At present, the time series prediction method based on data-driven modeling can be divided into global model and local model.The time series data obtained from the real world are often characterized by strong nonlinearity and uncertainty, which makes it difficult for the traditional prediction models to be effectively modeled, thus limiting the prediction accuracy of the models.Although the global model based on all historical data is simple, it is sensitive to the outliers in the sequence.In order to solve the above problems, a local modeling method based on cloud model similarity measurement is proposed in this paper, and a prediction model based on BP neural network (BP) and least squares support vector machine (LS-SVM) is established, which can effectively improve the prediction accuracy of the sequence.Although the local modeling methods mentioned above have a certain contribution to the prediction accuracy, they still belong to the category of single-task learning. The implicit correlation information in time series is not fully mined, thus affecting the generalization performance of the model.Considering the advantages of local model and multitask learning, a local modeling method based on multitask learning is proposed in this paper.In this method, the multiple information contained in the adjacent time points of time series can be regarded as different tasks in multitask learning, so as to improve the generalization ability of the model.The research work of this paper mainly includes the following aspects: 1) analyzing the advantages and disadvantages of several common time series similarity measurement methods.Two different kinds of time series prediction methods are compared. The original time series data are nonlinear, complex and uncertain.The cloud model theory is used to represent the time series of the original sequence and the first order difference processing sequence simultaneously. In view of the traditional distance function, it is difficult to measure the data containing uncertain factors effectively.This paper proposes a local modeling method based on similarity measurement of cloud model, and uses BPNN and LS-SVM to build prediction model for modeling and forecasting.In the framework of the proposed local modeling method, multitask learning is applied to time series prediction.A local modeling method based on multitasking learning is proposed. The proposed method is validated by six real engineering data sets selected from the Machine Learning Application Group of the School of Science of Aalto University.
【學(xué)位授予單位】：太原理工大學(xué)
【學(xué)位級(jí)別】：碩士
【學(xué)位授予年份】：2017
【分類號(hào)】：O211.61

【參考文獻(xiàn)】

相關(guān)期刊論文 前10條

1 邵蔚元;郭躍飛;;多任務(wù)學(xué)習(xí)及卷積神經(jīng)網(wǎng)絡(luò)在人臉識(shí)別中的應(yīng)用[J];計(jì)算機(jī)工程與應(yīng)用;2016年13期

2 孫妮妮;陳澤華;牛昱光;閻高偉;;基于云模型重疊度的相似性度量[J];計(jì)算機(jī)應(yīng)用;2015年07期

3 門連生;衛(wèi)婧菲;李中;;基于形態(tài)相似距離的時(shí)間序列相似性度量[J];計(jì)算機(jī)工程與應(yīng)用;2015年04期

4 金璐;覃思義;;基于云模型間貼近度的相似度量法[J];計(jì)算機(jī)應(yīng)用研究;2014年05期

5 劉成;彭進(jìn)業(yè);;基于多任務(wù)學(xué)習(xí)的自然圖像分類研究[J];計(jì)算機(jī)應(yīng)用研究;2012年07期

6 王新迎;韓敏;;基于極端學(xué)習(xí)機(jī)的多變量混沌時(shí)間序列預(yù)測(cè)[J];物理學(xué)報(bào);2012年08期

7 李海林;郭崇慧;邱望仁;;正態(tài)云模型相似度計(jì)算方法[J];電子學(xué)報(bào);2011年11期

8 張光衛(wèi);李德毅;李鵬;康建初;陳桂生;;基于云模型的協(xié)同過(guò)濾推薦算法[J];軟件學(xué)報(bào);2007年10期

9 曲文龍;張德政;楊炳儒;;基于小波和動(dòng)態(tài)時(shí)間彎曲的時(shí)間序列相似匹配[J];北京科技大學(xué)學(xué)報(bào);2006年04期

10 楊振成;自回歸滑動(dòng)平均模型中階數(shù)及參數(shù)的確定[J];統(tǒng)計(jì)與決策;2004年12期

相關(guān)碩士學(xué)位論文 前4條

1 黃杰;SOM在時(shí)間序列預(yù)測(cè)中的應(yīng)用研究[D];蘭州交通大學(xué);2015年

2 李楊;基于多任務(wù)學(xué)習(xí)的人臉圖像識(shí)別[D];上海大學(xué);2014年

3 劉凌波;基于多任務(wù)學(xué)習(xí)的年齡估計(jì)研究[D];中南大學(xué);2012年

4 金鳳;多視角的構(gòu)建及其在單任務(wù)學(xué)習(xí)和多任務(wù)學(xué)習(xí)中的應(yīng)用[D];華東師范大學(xué);2010年

，

本文編號(hào)：1722858