大连海洋大学学报
大連海洋大學學報
대련해양대학학보
JOURNAL OF DALIAN FISHERIES UNIVERSITY
2014年
5期
520-524
,共5页
多变量灰色系统%灰色关联度%MGM (1, n) 模型%微分方程组
多變量灰色繫統%灰色關聯度%MGM (1, n) 模型%微分方程組
다변량회색계통%회색관련도%MGM (1, n) 모형%미분방정조
multi-variable grey system%grey correlation degree%MGM(1,n) model%system of differential equa-tion
运用灰色关联度与多变量灰色系统理论对中国1986-2011年水产品产量时间序列进行了实证分析。首先用灰色关联分析方法将水产品产量所包含的5组时间序列即海洋捕捞、远洋渔业、海水养殖、淡水捕捞、淡水养殖产量进行了分类；再将关联度达到0.9以上的海洋捕捞、海水养殖、淡水养殖产量3组时间序列划分为一组并运用多变量灰色MGM (1, n)模型对其进行建模,由于这3组序列关联度高且单个序列累加数据近似服从指数分布,因而拟和结果数据逼近初始数据；对关联度达到0.8的远洋渔业与淡水捕捞两组时间序列,也尝试用MGM (1, n)模型进行建模,结果模型拟合略有欠缺。本研究中的这种建模方法既考虑了序列间的内在联系,又避免了分别使用单一序列建模割裂序列相互制约因素的情况发生,图示与误差分析结果均显示该方法具有效性。
運用灰色關聯度與多變量灰色繫統理論對中國1986-2011年水產品產量時間序列進行瞭實證分析。首先用灰色關聯分析方法將水產品產量所包含的5組時間序列即海洋捕撈、遠洋漁業、海水養殖、淡水捕撈、淡水養殖產量進行瞭分類；再將關聯度達到0.9以上的海洋捕撈、海水養殖、淡水養殖產量3組時間序列劃分為一組併運用多變量灰色MGM (1, n)模型對其進行建模,由于這3組序列關聯度高且單箇序列纍加數據近似服從指數分佈,因而擬和結果數據逼近初始數據；對關聯度達到0.8的遠洋漁業與淡水捕撈兩組時間序列,也嘗試用MGM (1, n)模型進行建模,結果模型擬閤略有欠缺。本研究中的這種建模方法既攷慮瞭序列間的內在聯繫,又避免瞭分彆使用單一序列建模割裂序列相互製約因素的情況髮生,圖示與誤差分析結果均顯示該方法具有效性。
운용회색관련도여다변량회색계통이론대중국1986-2011년수산품산량시간서렬진행료실증분석。수선용회색관련분석방법장수산품산량소포함적5조시간서렬즉해양포로、원양어업、해수양식、담수포로、담수양식산량진행료분류；재장관련도체도0.9이상적해양포로、해수양식、담수양식산량3조시간서렬화분위일조병운용다변량회색MGM (1, n)모형대기진행건모,유우저3조서렬관련도고차단개서렬루가수거근사복종지수분포,인이의화결과수거핍근초시수거；대관련도체도0.8적원양어업여담수포로량조시간서렬,야상시용MGM (1, n)모형진행건모,결과모형의합략유흠결。본연구중적저충건모방법기고필료서렬간적내재련계,우피면료분별사용단일서렬건모할렬서렬상호제약인소적정황발생,도시여오차분석결과균현시해방법구유효성。
The empirical analysis of time series of output of fish and fishery products in China from 1986 to 2011 was conducted using the theories of grey correlation degree and multi-variable grey system. Firstly, the five time series of the output including output of marine capture, output of distance fisheries, output of mariculture, output of freshwater capture and output of freshwater culture were classified by a method of grey correlation analysis. The three time series of output of marine capture, output of mariculture and output of freshwater culture with grey corre-lation degree of more than 0 . 9 were categoried into one group and fitted the three sets of time series by multi-varia-ble grey model MGM ( 1 , n ) . The fitting data were approximate to the initial data in the three time series apart whose correlation degrees were high and distributions of the accumulate data were of exponential function. The two time series of output of distance fisheries and the output of freshwater capture with correlation degree of more than 0. 8 were attempted to be modeled by MGM(1,n). The internal relation among the time series was taken into ac-count and the situation of splitting the mutual restricting factors in the time series was avoided when the models were reconstructed only by every time series separately. The effectiveness of the method is approved by graphics and the errors analysis.
