頁籤選單縮合
題名 | 應用複變值直線迴歸分析法研究水稻品種地域適應性 |
---|---|
作者姓名(中文) | 沈明來; 魏應澤; | 書刊名 | 中華農學會報 |
卷期 | 71 1970.09[民59.09] |
頁次 | 頁1-13 |
關鍵詞 | 水稻品種; 直線迴歸分析法; 複變值; |
語文 | 中文(Chinese) |
中文摘要 | (一)本研究為搜集臺灣省農業試驗所,臺中區農業改良場,嘉義農業試驗分所及高雄區農業改良場自民35年至54年20年間之水稻豐歉試驗之成績及氣象因素因素紀錄為資料。水稻農藝性狀採用生育日數(Y1)株高(Y2)一株穗數(Y3)及穀產量(Y4)等四種;氣象因素採用溫度(X1)日照(X2)濕度(X3)溫差(X4)雨量(X5)等五種。以農藝性狀為依變數(Dependent variables),氣象因素為自變數(Independent variables)。應用複變值直線?歸分析法(Multivariate linear regression analysis),以研究水稻品種地域適應性。 (二)複變值常態分佈β由單變值之聯合機率密度(Joint probability density) 誘導而成,設有P種逢機變值(Random variates),其數學模式為 Y=BX+d 其Y之聯合機率密度為 g(Y)= 1/(2π)?√|Σ| exp[- ?(Y-BX)'Σ?1(Y-BX)] (三)複變值常態分佈未知常數Σ及B可用最大機率推算法(Masimum likelihood estimation)求得 B=(X'X)?1(X'Y) Σ=1/n-q(Y'Y-B'X'XB) (四)測驗複變值?歸係數向量顯著性之統計值為x-q-p+1/p.T2/n-q,其分布為自由度等於p及x-q-p+1之F分布,式中T2為HOTELLING氏統計值,其一般公式為T2=d'Σ?1d。 (五)各種水稻品種對於各項氣象因子反應各異,一般來講,蓬萊稻對於氣象因子之反應較為遲鈍,而在來稻則較為敏感,尤其對於溫度及日照更為明顯。 (六)臺灣省北、中、南部氣候之變化迥異,其中以北部及南部氣候變化較大,中部較小。是以臺中65號在各地之生長發育亦有所差別。中部地域最適合其生長,南部次之,北部較差。 |
英文摘要 | 1. The identification of rice varieties adaptable to various climates and locations is discussed by using the multivariate linear regression analysis in which the agronomic characteristics of rice varieties, such as number of growth days after transplanting (Y1), plant height (Y2), number of pancles per hill (Y3), and grain yield (Y4) are used as regressands while the climate elements such as temperature (X1), hours of sunlight (X2), humidity (X3), temperature difference between day and night (X4), and rainfall (X5) are used as regressors. The data used in this study were collected from the results of experiments conducted by Taiwan Agricultural Experiment Stations. 2. The normal distribution of multivariate is derived from the joint probability density of single variate. The general mathematical model of p random variates is Y=BX+d and the joint probability density of Y is g(Y)= 1/(2π)?√|Σ| exp[- ?(Y-BX)'Σ?1(Y-BX)] 3. The following estimators for Σ and B are derived from the method of maximum likehood: B=(X'X)?1(X'Y) Σ=1/n-q(Y'Y-B'X'XB) 4. The hypothesis that a particular vector in B equals to zero is tested by the statistic x-q-p+1/p.T2/n-q, which has F distribution with p and x-q-p+1 degrees of freedom, where T2 is Hotelling's statistic. 5. The responses of rice varieties to climate elements are different. In general, the Indica rice is more sensitive to climate elements than Japonica rice, especially to temperature and sunlight. 6. The climate of both northern and southern parts of Taiwan is more variable than that of the central part. It is found that the central Taiwan is the most adaptable area for growing Taichung 65 of rice variety. |
本系統之摘要資訊系依該期刊論文摘要之資訊為主。