本文選題：復(fù)權(quán)馬爾可夫鏈 + 數(shù)值預(yù)測(cè)　； 參考：《南水北調(diào)與水利科技》2017年06期

【摘要】：針對(duì)傳統(tǒng)馬爾可夫鏈及其改進(jìn)的預(yù)測(cè)方法只能進(jìn)行狀態(tài)預(yù)測(cè)的局限,根據(jù)相依隨機(jī)變量的特點(diǎn),在以傳統(tǒng)馬爾可夫鏈預(yù)測(cè)方法求得各狀態(tài)預(yù)測(cè)概率的基礎(chǔ)上,進(jìn)一步以狀態(tài)預(yù)測(cè)概率為權(quán)重與狀態(tài)平均值加權(quán)求和,實(shí)現(xiàn)了馬爾可夫鏈預(yù)測(cè)方法從狀態(tài)預(yù)測(cè)到數(shù)值預(yù)測(cè)的關(guān)鍵性改進(jìn)。利用我國(guó)西南國(guó)際大河怒江干流道街壩水文站1957-2010年徑流和1964-2010年懸移質(zhì)輸沙序列為分析期,2011-2015年徑流和懸移質(zhì)輸沙為驗(yàn)證期,對(duì)所建立的復(fù)權(quán)馬爾可夫鏈預(yù)測(cè)方法步驟進(jìn)行驗(yàn)證表明,復(fù)權(quán)馬爾可夫鏈預(yù)測(cè)方法具有較高的數(shù)值預(yù)測(cè)精度,能夠滿足隨機(jī)時(shí)間序列短期數(shù)值預(yù)測(cè)的需要。
[Abstract]:According to the characteristics of the traditional Markov chain and its improved prediction method, according to the characteristics of the dependent random variables, the Markov chain preview is realized on the basis of the traditional Markov chain prediction method, and the state prediction probability is weighted and the state mean value is weighted. The method from state prediction to numerical prediction is a key improvement. The 1957-2010 year runoff and 1964-2010 year suspended sediment transport sequence in the main stream Road dam of Nu River in the southwest of our country are used as the analysis period, and the 2011-2015 year runoff and suspended sediment transport is the verification period. The results show that the prediction method of complex Markov chain has higher numerical prediction accuracy and can meet the needs of short-term numerical prediction of random time series.
【作者單位】： 河海大學(xué)水文水資源與水利工程科學(xué)國(guó)家重點(diǎn)實(shí)驗(yàn)室;云南省滇中引水工程建設(shè)管理局昆明分局;云南財(cái)經(jīng)大學(xué)國(guó)土資源與持續(xù)發(fā)展研究所;云南省水文水資源局;
【基金】：江蘇省博士后科研資助計(jì)劃(1501060B) 云南省水利科技項(xiàng)目(2014003)~~
【分類號(hào)】：P333
，

本文編號(hào)：2059465