課程名稱︰統計學與實習下
課程性質︰系訂必修
課程教師︰林惠玲
開課系所︰經濟系
考試時間︰2006/06/06
試題 :
考試時間:2:20PM~5:00PM
一、(20%)單選題(無需理由)
Given the OLS regression results from a sample from t=1 to t=T,
^
y_t = 92.05 + 0.142x_1t + 0.236x_2t R^2 = 0.81
(11.85) (0.16) (0.131)
the values in parentheses are standard deviations.
Answer the question 1 to 6
1.Based on the estimated results, what kind of data problem you will suspect?
(A) Autocorrelation
(B) Heteroskedasticity
(C) Multicollinearity
2.Based on the estimated results,
(A) variable x_1 has a larger effect on y than variable x_2
(B) variable x_1 has a smaller effect on y than variable x_2
(C) variables x_1 and x_2 have the same effect on y
(D) the effects are not comparable
3.If the variable x_1 is dropped from the regression, the coefficient of
determination (R^2) will be
(A) increased
(B) decreased
(C) not changed
(D) no way to determine
4.If the variable x_1 is dropped from the regression, the adjusted
_
coefficient of determination (R^2 ) will be
(A) increased
(B) decreased
(C) not changed
(D) no way to determine
5. Denote R_1^2 as the coefficient of determination of the new regression
of log(y_t) on x_1t and x_2t. Suppose R_1^2 is greater than 0.81, the
coefficient of determination from the original regression. Then, we
can conclude that
(A) the new model is better than the original model
(B) the new model is worse than the original model
(C) no way to make a comparison
(D) two models have equal performance
6.If two variable are correlated by the least-squares technique and r=0.70,
which of the following is most correct?
(A) the percentage of squared deviations in the y values that are
"explained" by the x values is 30%
(B) the percentage of squared deviations in the y values that are
"explained" by the x values is 70%
(C) the percentage of squared deviations in the y values that are not
"explained" by the x values is 9%
(D) the percentage of squared deviations in the y values that are not
"explained" by the x values is 51%
(E) None of the above
二、(28%)某工管人員研究握力 (X) 與抬舉重量 (Y)(以公斤衝量)之間的關係。他找了20
位受測者蒐集資料,並將結果整理如下:
n = 20,ΣX = 550,ΣY = 1620,ΣX^2 = 16080,ΣY^2 = 135000,ΣXY = 45900
1. 試求X與Y間的相關係數 r,並檢定是否顯著 (α=0.05)。(7%)
2. 試求X預測Y的線性迴歸,並檢定其顯著性 (α=0.05)。(7%)
3. 若Y的衡量尺度以十公斤來衡量,則Y對X估計的迴歸參數(截距項與斜率)為何?
與(2)結果相同嗎? (5%)
4. 承3小題,相關係數為何?會相同嗎? (4%)
5. 承3小題,1、2的檢定結果是否會不同?理由。(5%)
三、(20%)下列為家庭的食物支出、所得、人口數之資料:
食物支出(Y) 40 50 50 70 80 100 110 140 160 200
所得(X1) 100 150 100 300 200 400 300 300 850 300
人口數(X2) 1 1 2 1 4 2 4 6 1 8
已知 ^ ^ ^ ^
1: y = α + β_1X_1 + β_2X_2 = 0.56 + 0.167X_1 + 16.4X_2
^
ΣY^2 = 24253 ΣY^2 = 225200 Σe^2 = 947
請回答下列問題:
(1)、利用OLS估計 Y = B_0 + B_y1X_1 + ε_2 ;Y = c_0 + c_y2X_2 + ε_3 。
^ ^ ^
(2)、請比較β_1,b_y1,c_y2 之關係,並說明其意義。
_
(3)、求1中對應的 R^2 及 R^2
(4)、檢定H_0 : β_1 = β_2 = 0 ; H_1 : H_0不為真(α=5%)
四、(24%)依據一組樣本數為45 的樣本資料進行迴歸分析,得到下列的結果
Y_i = 21000 + 3600X_1i + 5800X_2i R^2 = 0.78
(2121.7) (365.7) (562.3)
其中 Y_i = 年收入,以元為單位;
X_1i =1 學歷為大學, =0 其他 ;
X_2i =1 學歷為碩士, =0 其他 ;
X_3i =1 學歷為高中, =0 其他 ;
括弧內數值為估計標準誤。試回答下列問題:
1. 為何迴歸式未包含變數 X_3i? (3%)
2. 解釋 R^2 = 0.78的意義為何? (3%)
3. 請說明截距項係數之意義 (3%)
4. 請說明解釋變數 X_1i 與變數 X_2i 之迴歸係數的意義 (5%)
5. 請在5%顯著水準下,檢定學歷是否對於年收入有影響? (5%)
6. 如果以變數 X_3i 代替變數 X_1i 重新估計迴歸式,則新的迴歸式為何?(5%)
五、(16%)設食物支出( )與所得( )的迴歸模型為如下:
1:Y = α + βX + ε
2:Y = α + β(1/X) + ε
3:Y = (αe^βX)ε
4:Y = (εαX^β)
請分別說明上式各 β 之涵意,並簡單說明是否均可以OLS估計未知參數。
--
陷自己於忙碌之中 也許不寂寞但卻很孤獨......
陪伴的人也許很多 不孤獨卻不代表不寂寞......
--
※ 發信站: 批踢踢實業坊(ptt.cc)
◆ From: 140.112.251.145