Yahoo奇摩知識+ 將於 2021 年 5 月 4 日 (美國東部時間) 終止服務。自 2021 年 4 月 20 日 (美國東部時間) 起，Yahoo奇摩知識+ 網站將會轉為唯讀模式。其他 Yahoo奇摩產品與服務或您的 Yahoo奇摩帳號都不會受影響。如需關於 Yahoo奇摩知識+ 停止服務以及下載您個人資料的資訊，請參閱說明網頁。

# 有人可以幫忙以下的文章嗎 我想了解文章的意思

the within-meal carryover habit coefficient，and a random effect that represents the deviation of the nutrient's actual consumption from the consumption predicted by the level-1 model．In every level-2 equation for each panelist, a level-1 coefficient is modeled as a function of an intercept，two fixed effects，and a random effect to capture differences within the panelist's three meal-level coefficients．The meal dummy variables allow us to test for differences between meals in terms of carryover and baseline habits．Each level-3 equation includes an aggregate intercept and a random panelist effect．We assume that the error terms from different levels are independent. Since the level-1 model has four coefficients，we include only those panelists for whom there are five or more observations for each meal occasion. In the equations, the fixed effects are represented by theδ's, and the random effects are represented by the ε’s，μ’s , andυ’s．Theυandμ random effects enable us to control for panelist effects and estimate the extent of cross-panelist heterogeneity in habit coefficients. This is achieved by allowing each panelist to have a different deviation around each aggregate habit coefficient (δ)．The three types of random effects，one type corresponding to each level, cumulatively reflect the total variance in a dependent variable. If allμ p.m.i＝0 and all υ＝0, then the HLM model reduces to a regular OLS regression model. We mean-centered our three independent variables within each panelist. This enables us to interpret the intercepts as baseline or mean consumption levels．We test for base-line habit via the three aggregate or level-3 level baselines：breakfast，lunch，and dinner．We test for carryover habit via the three aggregate or level-3 lag estimates．Using breakfast as an example, these areδ1,0,0 andδ2,0,0 as the first and second lag across-meal carryover habit estimates and δ3,0,0 as the within-meal carryover habit estimate.

### 1 個解答

• 1 0 年前
最佳解答

吃飯中的轉入蒙上，並且係數，從消費表示營養物的現實消費的越出的隨機的效果為了各panelist水平-1model.為了預測了，並且由於In的所有水平-2方程式，水平-1個係數，在panelist的3個吃飯水平係數中得差異是不是沒有intercept的機能，2個固定效果，波及的作為隨機的效果被模型化做的;吃飯的差異以假裝遞球變量轉入和基線habits.用Each水平-3方程式試驗我們的吃飯來自以獨立的狀態不同的水平的誤差項在的集合intercept和effect.因為含著We假定的隨機的panelist的; 水平-1個模型有4個係數，我們只每把各自的吃飯的時候有5個以上的觀測的那些的panelist作為。 有效為了用方程式，固定效果根據δ被表示，隨機的效果根據ε被表示，microcosm的東西，υ非人為的效果，我們panelist效果和交叉的panelist的異種性的範圍在中蒙上係數這樣的估計的控制．Theυ和microcosm是那樣。 集合習慣係數(δ)．根據The3類型隨機的效果各自周圍不同的越出容許各panelist有的事，這個被實現，為了各自積累性地很平對應的1個類型從變量反映全分散。 全部微小，下午，i=0全部的υ=0，當時的HLM模型減少到通常的OLS回歸模型。 我們，平均中心，各panelist中的我們的3個自變量。 基線使壞的consumption levels.We是不是沒有跑壘道習慣用3集合試驗，或水平-3很平基線的時候，這，我們使解釋intercept成為可能的第:1號和秒作為吃飯中的轉入習慣估計與吃飯的對面的轉入習慣估計使之遲到δ3，0，0的時候，這些，通過3集合或者水平-3的轉入為了習慣的早飯，午飯，和dinner.We試驗作為例子estimates.使之遲到Using早飯，是δ1，0，0和δ2，0，0。

參考資料： 翻譯網