1. NCCU Academic Hub
  2. Publications
  3. College of Commerce
  4. Theses
  5. Item

Publications-Theses

Article View/Open

Publication Export

Google ScholarTM

NCCU Library

Citation Infomation

Related Publications in TAIR

題名 時間數列分析在偵測型態結構差異上之探討
Application Of Time Series Analysis In Pattern Recgnition And alysis
作者 蘇曉楓
Su, Shiau Feng
貢獻者 吳柏林
Wu, Berlin
蘇曉楓
Su, Shiau Feng
關鍵詞 非線性時間數列模式
神經網路
穩健性
模型辨識
時間數列分析
nonlinear time series
neural
time series analysis
日期 1993
上傳時間 29-Apr-2016 16:43:56 (UTC+8)
摘要 依時間順序出現之一連串觀測值，通常會呈現某一型態，而根據所產生的型態可以作為判斷事件發生的基礎。例如，震波形成原因的判斷﹔追查環境污染源﹔以及在醫學方面，辨識一個正常人心電圖的型態與患有心臟病的病人其心電圖的型態…等。對於這些問題，傳統之辨識方法常因前提假設的限制而失去其準確性。在本文中，我們應用神經網路中的逆向傳播演算法則來訓練網路，並利用此受過訓練的網路來辨別線性時間數列ARIMA及非線性時間數列 BL(1,0,1,1)。結果發現，網路對於模擬資料中雙線性係數介於0.2至$0.8$之間的資料有高達$80\\%$以上的辨識能力。而在實例研究中，我們訓練網路來判斷震波形成的原因，其正確率亦高達80\\%以上。同時，我們也將神經網路應用在環境保護方面，訓練網路來判斷二地區空氣品質的型態。
A series of observations indexed in time often produces a
參考文獻 Brockett,R.W. (1976). Volterra series and geometric control theory, A`utomatica
     ,Vo1.12,167-176.
     Chin,L. C. (1985). vVhat is biostatistics? ,Biometrics,41 ,771-775.
     Chan,D.Y.C. & Prager,D.(1991).Analysis of time series by neural networks:
     IEEE,355-360.
     De Gooi.jer,J.G.& Kumar,K.(1992).Some recent developments in nonlinear
     time series lllodelling,testing and forecasting.International Journal of
     Forecasting ,8,135-156.
     Farrugia,S., Yee,H.& Nickolls,P.,( 1991) .Neural networks classification of intracardiac
     ECGS,JEEE~1278-1283.
     Gedeon,T.D. & Harris,D.(1991).Creating robust networks,JEEE, 2553-2557.
     Ghosh,J.) Deuser) L. wI. & Beck,S. D. (1992). A nerual network based
     hybird systen1 for detection, characterization , and classification of
     short-duration oceanic signals, IEEE Journal Of Oceanic Engineer`
     lng ,Vo1.17 No.4 October , 351-363.
     Gorman,R.P. & Seinowski,T.J. (1988).Analysis of hidden units 1Il a layered
     netwok trained to classify sonar targets,Neural Networks,Vo1.1,75-89.
     Granger,C.vV.J.& Anderson)\\.P.(1978).An Introd`uction to Bilinear Ti`me Series
     M odels,Vandenhoeck and Ruprecht,Gottingent.
     Granger,C."\\;V.J.(1991).Developments in the nonlinear analysis of econoillic
     series.Scand.1. of Econo`mics,93(2) ,263-276.
     Guegan,D & Phalll,T.D.(1992).Power of the score test against bilinear tilDe
     series models.Statistica Sinica,Vo1.2))57-169.
     Kanaya,F.& IVIiyake,S.(1991).Bayes statistical behavior and valid generalization
     of pattern classifying neural networks, IEEE transactio`f1 on N e`ural
     lVetworks,Vo1.2,No.4,J uly,4 71-475.
     Ljung,G.:NI. (1978). On a lneasure of lack of fit In tilne senes l1lodels
     Bio`me tries, Vol. 65,297 -303.
     Lipplllann,R .P. (1989). Pattern classification uSIng neural net-
     works IEEE Corn;munication Magazine.
     Mohler ,R. R. (1973). Bilinear control processes ,Acaclenlic Press, New Yorkand London.
     Robert,J.S.(1992).Pattern Recognition.
     Ruberti,A.,Isidori A. & d`Allessanclro,P.(1972). Theory of bilinear dynam,ical
     system,Springer Verlag,Berlin.
     Shulllway,R.H.(1988).Applied Statistical Ti`me Series Ana.lysis.
     Shibata,R.(1976). Selection of the order of an autoregressive nlodel by
     akaike`s infonnation criterion ,Biometrics, Vol.63(1) , 117.
     Saikkonen,P.& Luukkonen,R.(1988). Lagrange multiplier tests for testing
     nonlineari ties in time series lllodeis ,S cand J oural of Statistics, 55-68.
     Saikkonen, P. & Luukkonen,R. (1988).Power properties of a tilne series linearity
     test against SOllle simpe bilinear alternatives,Statistica Sinica,453-464.
     Subba Rao,T. & Gabr,NI.NI.(1984).An Introduction to Bispectral Analysis and
     Bilinear Time Series Nlodels;Springer- Verlag,Berlin.
     Terence, C.NI. (1992) .NIodelling the seasonal patterns in UI( macroeconomic
     tilnes series,Journal of Royal Statistical Society ,61-75.
     Tsay,R.S. (1991) .Detecting and modeling nonlinearity in univariate tinle series
     analysis,Statistica Sinica,431-451.
     Takayuki, Y.,Tetsuro, Y.(1992).Neural networks controller USlllg autotun-ing
     methodfor nonlinear function,IEEE Transcation on N ev:ral JVetworks,
     3( 4), 595-601.
     Tong,H.& Lim,I(.S. (1980).Tlueshold autoregressioIl,limit cycles and cyclical
     data. l.Roy. Statist. Soc.Ser.B,42)45-292.
     \\iVeigend,A.S.& Rumelhart,D.E.(1991).The effective dilnention of space of
     hidden U nits,IEEE,2069-207 4.
     Yee E. & Ho J.( 1990). Neural nenwork recognition And classification of
     aecrosol distri butions nleasured with a two-spot laser velocimeter ,A pplied
     Opiics ,2929-2938.
     Zaknich,A. & Attikiouzel,Y. (1991) . A 1110dified probabilistic neural network
     (PNN) for nonlinear ti111e series analysis, IEEE ,1530-1535.
描述 碩士
國立政治大學
統計學系
G80354009
資料來源 http://thesis.lib.nccu.edu.tw/record/#B2002004195
資料類型 thesis
dc.contributor.advisor 吳柏林zh_TW
dc.contributor.advisor Wu, Berlinen_US
dc.contributor.author (Authors) 蘇曉楓zh_TW
dc.contributor.author (Authors) Su, Shiau Fengen_US
dc.creator (作者) 蘇曉楓zh_TW
dc.creator (作者) Su, Shiau Fengen_US
dc.date (日期) 1993en_US
dc.date.accessioned 29-Apr-2016 16:43:56 (UTC+8)-
dc.date.available 29-Apr-2016 16:43:56 (UTC+8)-
dc.date.issued (上傳時間) 29-Apr-2016 16:43:56 (UTC+8)-
dc.identifier (Other Identifiers) B2002004195en_US
dc.identifier.uri (URI) http://nccur.lib.nccu.edu.tw/handle/140.119/89020-
dc.description (描述) 碩士zh_TW
dc.description (描述) 國立政治大學zh_TW
dc.description (描述) 統計學系zh_TW
dc.description (描述) G80354009zh_TW
dc.description.abstract (摘要) 依時間順序出現之一連串觀測值，通常會呈現某一型態，而根據所產生的型態可以作為判斷事件發生的基礎。例如，震波形成原因的判斷﹔追查環境污染源﹔以及在醫學方面，辨識一個正常人心電圖的型態與患有心臟病的病人其心電圖的型態…等。對於這些問題，傳統之辨識方法常因前提假設的限制而失去其準確性。在本文中，我們應用神經網路中的逆向傳播演算法則來訓練網路，並利用此受過訓練的網路來辨別線性時間數列ARIMA及非線性時間數列 BL(1,0,1,1)。結果發現，網路對於模擬資料中雙線性係數介於0.2至$0.8$之間的資料有高達$80\\%$以上的辨識能力。而在實例研究中，我們訓練網路來判斷震波形成的原因，其正確率亦高達80\\%以上。同時，我們也將神經網路應用在環境保護方面，訓練網路來判斷二地區空氣品質的型態。zh_TW
dc.description.abstract (摘要) A series of observations indexed in time often produces aen_US
dc.description.tableofcontents 壹　前言‧‧‧‧‧‧‧‧‧‧‧‧‧‧‧‧‧‧‧‧‧‧‧‧‧‧‧‧‧‧　 ３
     
     貳　型態辨識探討‧‧‧‧‧‧‧‧‧‧‧‧‧‧‧‧‧‧‧‧‧‧‧‧‧‧　 ６
     　2.1　型態辨識的方法‧‧‧‧‧‧‧‧‧‧‧‧‧‧‧‧‧‧‧‧‧‧‧　 ７
      2.1.1　靜態資料‧‧‧‧‧‧‧‧‧‧‧‧‧‧‧‧‧‧‧‧‧‧‧‧　 ７
      2.1.2　動態資料‧‧‧‧‧‧‧‧‧‧‧‧‧‧‧‧‧‧‧‧‧‧‧　１１
      2.1.3　穩健性的探討‧‧‧‧‧‧‧‧‧‧‧‧‧‧‧‧‧‧‧‧‧　１２
     
     參　神經網路在非線性時間數列模型辨識之應用
     　3.1　神經網路介紹‧‧‧‧‧‧‧‧‧‧‧‧‧‧‧‧‧‧‧‧‧‧　 １４
     　3.2　雙線性模式的探討‧‧‧‧‧‧‧‧‧‧‧‧‧‧‧‧‧‧‧‧　 １７
     　3.3　應用神經網路做模型辨識‧‧‧‧‧‧‧‧‧‧‧‧‧‧‧‧‧　 ２０
     　3.1　模擬比較與結果‧‧‧‧‧‧‧‧‧‧‧‧‧‧‧‧‧‧‧‧‧　 ２２
     
     肆　實例研究
     　例4.1　地震震波與核子試爆震波的辨識‧‧‧‧‧‧‧‧‧‧‧‧‧　 ２９
     　例4.2　環保污染品質型態的判別‧‧‧‧‧‧‧‧‧‧‧‧‧‧‧‧ 　３０
     
     伍　結論‧‧‧‧‧‧‧‧‧‧‧‧‧‧‧‧‧‧‧‧‧‧‧‧‧‧‧‧　３４
     
     參考文獻‧‧‧‧‧‧‧‧‧‧‧‧‧‧‧‧‧‧‧‧‧‧‧‧‧‧‧‧　３５		zh_TW
dc.source.uri (資料來源) http://thesis.lib.nccu.edu.tw/record/#B2002004195en_US
dc.subject (關鍵詞) 非線性時間數列模式zh_TW
dc.subject (關鍵詞) 神經網路zh_TW
dc.subject (關鍵詞) 穩健性zh_TW
dc.subject (關鍵詞) 模型辨識zh_TW
dc.subject (關鍵詞) 時間數列分析zh_TW
dc.subject (關鍵詞) nonlinear time seriesen_US
dc.subject (關鍵詞) neuralen_US
dc.subject (關鍵詞) time series analysisen_US
dc.title (題名) 時間數列分析在偵測型態結構差異上之探討zh_TW
dc.title (題名) Application Of Time Series Analysis In Pattern Recgnition And alysisen_US
dc.type (資料類型) thesisen_US
dc.relation.reference (參考文獻) Brockett,R.W. (1976). Volterra series and geometric control theory, A`utomatica
     ,Vo1.12,167-176.
     Chin,L. C. (1985). vVhat is biostatistics? ,Biometrics,41 ,771-775.
     Chan,D.Y.C. & Prager,D.(1991).Analysis of time series by neural networks:
     IEEE,355-360.
     De Gooi.jer,J.G.& Kumar,K.(1992).Some recent developments in nonlinear
     time series lllodelling,testing and forecasting.International Journal of
     Forecasting ,8,135-156.
     Farrugia,S., Yee,H.& Nickolls,P.,( 1991) .Neural networks classification of intracardiac
     ECGS,JEEE~1278-1283.
     Gedeon,T.D. & Harris,D.(1991).Creating robust networks,JEEE, 2553-2557.
     Ghosh,J.) Deuser) L. wI. & Beck,S. D. (1992). A nerual network based
     hybird systen1 for detection, characterization , and classification of
     short-duration oceanic signals, IEEE Journal Of Oceanic Engineer`
     lng ,Vo1.17 No.4 October , 351-363.
     Gorman,R.P. & Seinowski,T.J. (1988).Analysis of hidden units 1Il a layered
     netwok trained to classify sonar targets,Neural Networks,Vo1.1,75-89.
     Granger,C.vV.J.& Anderson)\\.P.(1978).An Introd`uction to Bilinear Ti`me Series
     M odels,Vandenhoeck and Ruprecht,Gottingent.
     Granger,C."\\;V.J.(1991).Developments in the nonlinear analysis of econoillic
     series.Scand.1. of Econo`mics,93(2) ,263-276.
     Guegan,D & Phalll,T.D.(1992).Power of the score test against bilinear tilDe
     series models.Statistica Sinica,Vo1.2))57-169.
     Kanaya,F.& IVIiyake,S.(1991).Bayes statistical behavior and valid generalization
     of pattern classifying neural networks, IEEE transactio`f1 on N e`ural
     lVetworks,Vo1.2,No.4,J uly,4 71-475.
     Ljung,G.:NI. (1978). On a lneasure of lack of fit In tilne senes l1lodels
     Bio`me tries, Vol. 65,297 -303.
     Lipplllann,R .P. (1989). Pattern classification uSIng neural net-
     works IEEE Corn;munication Magazine.
     Mohler ,R. R. (1973). Bilinear control processes ,Acaclenlic Press, New Yorkand London.
     Robert,J.S.(1992).Pattern Recognition.
     Ruberti,A.,Isidori A. & d`Allessanclro,P.(1972). Theory of bilinear dynam,ical
     system,Springer Verlag,Berlin.
     Shulllway,R.H.(1988).Applied Statistical Ti`me Series Ana.lysis.
     Shibata,R.(1976). Selection of the order of an autoregressive nlodel by
     akaike`s infonnation criterion ,Biometrics, Vol.63(1) , 117.
     Saikkonen,P.& Luukkonen,R.(1988). Lagrange multiplier tests for testing
     nonlineari ties in time series lllodeis ,S cand J oural of Statistics, 55-68.
     Saikkonen, P. & Luukkonen,R. (1988).Power properties of a tilne series linearity
     test against SOllle simpe bilinear alternatives,Statistica Sinica,453-464.
     Subba Rao,T. & Gabr,NI.NI.(1984).An Introduction to Bispectral Analysis and
     Bilinear Time Series Nlodels;Springer- Verlag,Berlin.
     Terence, C.NI. (1992) .NIodelling the seasonal patterns in UI( macroeconomic
     tilnes series,Journal of Royal Statistical Society ,61-75.
     Tsay,R.S. (1991) .Detecting and modeling nonlinearity in univariate tinle series
     analysis,Statistica Sinica,431-451.
     Takayuki, Y.,Tetsuro, Y.(1992).Neural networks controller USlllg autotun-ing
     methodfor nonlinear function,IEEE Transcation on N ev:ral JVetworks,
     3( 4), 595-601.
     Tong,H.& Lim,I(.S. (1980).Tlueshold autoregressioIl,limit cycles and cyclical
     data. l.Roy. Statist. Soc.Ser.B,42)45-292.
     \\iVeigend,A.S.& Rumelhart,D.E.(1991).The effective dilnention of space of
     hidden U nits,IEEE,2069-207 4.
     Yee E. & Ho J.( 1990). Neural nenwork recognition And classification of
     aecrosol distri butions nleasured with a two-spot laser velocimeter ,A pplied
     Opiics ,2929-2938.
     Zaknich,A. & Attikiouzel,Y. (1991) . A 1110dified probabilistic neural network
     (PNN) for nonlinear ti111e series analysis, IEEE ,1530-1535.		zh_TW