112 年 國立政治大學財政學系碩士班在職生《統計學》

📄 試題原卷 免費註冊後即可對照原始考卷 PDF免費註冊

第 1 題20 分

The table below shows the winning percentages for 5 professional baseball teams during the 2021 and 2022 season.
Let X denote the winning percentage in 2022 and Y denote the percentage in 2021.

Teams2022 (X)2021 (Y)
Monkey0.6030.479
Elephant0.5950.574
Snake0.4960.427
Lions0.4100.557
Fighter0.3970.466

a) Find E(X+Y2)E\left(\frac{X+Y}{2}\right) and V(X+Y2)V\left(\frac{X+Y}{2}\right).
b) What is the correlation coefficient between the 2021 and 2022 season winning percentages?

登入後即可作答並保存紀錄。

這一題的完整詳解

本題主要在考驗對隨機變數期望值與變異數性質的理解,以及相關係數的計算。

核心觀念:

  1. 期望值性質: 對於任意常數 a,ba, b 和隨機變數 X,YX, Y,有 E(aX+bY)=aE(X)+bE(Y)E(aX + bY) = aE(X) + bE(Y)。
  2. 變異數性質: 對於任意常數 a,ba, b 和隨機變數 X,YX, Y,若 X,YX, Y 獨立,則 V(aX+bY)=a2V(X)+b2V(Y)V(aX + bY) = a^2V(X) + b^2V(Y)。若 X,YX, Y 不獨立,則需考慮共變異數 V(aX+bY)=a2V(X)+b2V(Y)+2abCov(X,Y)V(aX + bY) = a^2V(X) + b^2V(Y) + 2ab Cov(X, Y)。
  3. 相關係數: 衡量兩個隨機變數之間線性相關的強度與方向,計算公式為 ρXY=Cov(X,Y)V(X)V(Y)\rho_{XY} = \frac{Cov(X, Y)}{\sqrt{V(X)V(Y)}}。
  4. 樣本統計量: 在此題中,我們將樣本數據視為對應隨機變數的實現值,並計算樣本均值、樣本變異數等作為對應期望值與變異數的估計。

解題過程:

首先,我們將給定的數據整理成表格,並計算相關的樣本統計量。
令 XiX_i 為第 ii 個隊伍在 2022 年的勝率,YiY_i 為第 ii 個隊伍在 2021 年的勝率。
數據如下:

隊伍XiX_i (2022)YiY_i (2021)
Monkey0.6030.479
Elephant0.5950.574
Snake0.4960.427
Lions0.4100.557
Fighter0.3970.466

樣本數 n=5n=5。

計算樣本均值:
Xˉ=1n∑i=1nXi=0.603+0.595+0.496+0.410+0.3975=2.5015=0.5002\bar{X} = \frac{1}{n}\sum_{i=1}^n X_i = \frac{0.603 + 0.595 + 0.496 + 0.410 + 0.397}{5} = \frac{2.501}{5} = 0.5002
Yˉ=1n∑i=1nYi=0.479+0.574+0.427+0.557+0.4665=2.5035=0.5006\bar{Y} = \frac{1}{n}\sum_{i=1}^n Y_i = \frac{0.479 + 0.574 + 0.427 + 0.557 + 0.466}{5} = \frac{2.503}{5} = 0.5006

計算樣本變異數 (使用 n−1n-1 作為分母,即樣本變異數的標準定義):
SX2=1n−1∑i=1n(Xi−Xˉ)2S_X^2 = \frac{1}{n-1}\sum_{i=1}^n (X_i - \bar{X})^2
SY2=1n−1∑i=1n(Yi−Yˉ)2S_Y^2 = \frac{1}{n-1}\sum_{i=1}^n (Y_i - \bar{Y})^2

∑(Xi−Xˉ)2=(0.603−0.5002)2+(0.595−0.5002)2+(0.496−0.5002)2+(0.410−0.5002)2+(0.397−0.5002)2\sum (X_i - \bar{X})^2 = (0.603-0.5002)^2 + (0.595-0.5002)^2 + (0.496-0.5002)^2 + (0.410-0.5002)^2 + (0.397-0.5002)^2
=(0.1028)2+(0.0948)2+(−0.0042)2+(−0.0902)2+(−0.1032)2= (0.1028)^2 + (0.0948)^2 + (-0.0042)^2 + (-0.0902)^2 + (-0.1032)^2
=0.01056784+0.00898704+0.00001764+0.00813604+0.01065024=0.0383588= 0.01056784 + 0.00898704 + 0.00001764 + 0.00813604 + 0.01065024 = 0.0383588
SX2=0.03835885−1=0.03835884≈0.0095897S_X^2 = \frac{0.0383588}{5-1} = \frac{0.0383588}{4} \approx 0.0095897

∑(Yi−Yˉ)2=(0.479−0.5006)2+(0.574−0.5006)2+(0.427−0.5006)2+(0.557−0.5006)2+(0.466−0.5006)2\sum (Y_i - \bar{Y})^2 = (0.479-0.5006)^2 + (0.574-0.5006)^2 + (0.427-0.5006)^2 + (0.557-0.5006)^2 + (0.466-0.5006)^2
=(−0.0216)2+(0.0734)2+(−0.0736)2+(0.0564)2+(−0.0346)2= (-0.0216)^2 + (0.0734)^2 + (-0.0736)^2 + (0.0564)^2 + (-0.0346)^2
=0.00046656+0.00538756+0.00541696+0.00318096+0.00119716=0.0156492= 0.00046656 + 0.00538756 + 0.00541696 + 0.00318096 + 0.00119716 = 0.0156492
SY2=0.01564925−1=0.01564924≈0.0039123S_Y^2 = \frac{0.0156492}{5-1} = \frac{0.0156492}{4} \approx 0.0039123

a) Find E(X+Y2)E\left(\frac{X+Y}{2}\right) and V(X+Y2)V\left(\frac{X+Y}{2}\right)

假設 XX 和 YY 分別代表所有隊伍在 2022 年和 2021 年勝率的隨機變數。在此,我們使用樣本的期望值和變異數來估計總體的期望值和變異數。
由於我們只有 5 個數據點,我們將使用樣本均值 Xˉ\bar{X} 和 Yˉ\bar{Y} 作為 E(X)E(X) 和 E(Y)E(Y) 的估計,而樣本變異數 SX2S_X^2 和 SY2S_Y^2 作為 V(X)V(X) 和 V(Y)V(Y) 的估計。

期望值:
E(X+Y2)=12E(X+Y)=12(E(X)+E(Y))E\left(\frac{X+Y}{2}\right) = \frac{1}{2} E(X+Y) = \frac{1}{2} (E(X) + E(Y))
我們使用樣本均值來估計:
E(X+Y2)≈Xˉ+Yˉ2=0.5002+0.50062=1.00082=0.5004E\left(\frac{X+Y}{2}\right) \approx \frac{\bar{X}+\bar{Y}}{2} = \frac{0.5002 + 0.5006}{2} = \frac{1.0008}{2} = 0.5004

變異數:
V(X+Y2)=V(12X+12Y)=(12)2V(X)+(12)2V(Y)+2(12)(12)Cov(X,Y)V\left(\frac{X+Y}{2}\right) = V\left(\frac{1}{2}X + \frac{1}{2}Y\right) = \left(\frac{1}{2}\right)^2 V(X) + \left(\frac{1}{2}\right)^2 V(Y) + 2\left(\frac{1}{2}\right)\left(\frac{1}{2}\right) Cov(X, Y)
V(X+Y2)=14V(X)+14V(Y)+12Cov(X,Y)V\left(\frac{X+Y}{2}\right) = \frac{1}{4} V(X) + \frac{1}{4} V(Y) + \frac{1}{2} Cov(X, Y)

在此,我們必須計算樣本的共變異數 Cov(X,Y)Cov(X, Y):
Cov(X,Y)=1n−1∑i=1n(Xi−Xˉ)(Yi−Yˉ)Cov(X, Y) = \frac{1}{n-1}\sum_{i=1}^n (X_i - \bar{X})(Y_i - \bar{Y})
∑(Xi−Xˉ)(Yi−Yˉ)=(0.603−0.5002)(0.479−0.5006)+(0.595−0.5002)(0.574−0.5006)+(0.496−0.5002)(0.427−0.5006)+(0.410−0.5002)(0.557−0.5006)+(0.397−0.5002)(0.466−0.5006)\sum (X_i - \bar{X})(Y_i - \bar{Y}) = (0.603-0.5002)(0.479-0.5006) + (0.595-0.5002)(0.574-0.5006) + (0.496-0.5002)(0.427-0.5006) + (0.410-0.5002)(0.557-0.5006) + (0.397-0.5002)(0.466-0.5006)
=(0.1028)(−0.0216)+(0.0948)(0.0734)+(−0.0042)(−0.0736)+(−0.0902)(0.0564)+(−0.1032)(−0.0346)= (0.1028)(-0.0216) + (0.0948)(0.0734) + (-0.0042)(-0.0736) + (-0.0902)(0.0564) + (-0.1032)(-0.0346)
=−0.00222048+0.00695072+0.00030912+−0.00508728+0.00357072=0.0035228= -0.00222048 + 0.00695072 + 0.00030912 + -0.00508728 + 0.00357072 = 0.0035228
Cov(X,Y)=0.00352285−1=0.00352284≈0.0008807Cov(X, Y) = \frac{0.0035228}{5-1} = \frac{0.0035228}{4} \approx 0.0008807

🔒

後續完整解題步驟與【答案】

免費註冊,享三天全站完整詳解閱覽。

免費註冊

第 2 題15 分

A poll was taken this year asking college students if they support the bill to raise minimum wage by NT$2,000. A similar poll was taken five years ago. Results are summarized below.

Sample SizeNumber of Responses in Support of the Bill
Present Sample300150
Previous sample275121

a) Find the 99% confidence interval for the current support of the bill.
b) Suppose the researcher wants to reduce the margin of error of the 95% confidence interval to be within ±2\pm 2 percentage points. How large a sample should be taken the next time when she conducts the survey?
c) At 5% significance level, test whether the support for the bill has increased in the past 5 years.

登入後即可作答並保存紀錄。

這一題的完整詳解

本題主要考驗對比例的信賴區間估計、樣本大小的決定,以及比例的假設檢定。

核心觀念:

  1. 比例的信賴區間: 對於一個比例 pp,其信賴區間通常使用常態近似,公式為 p^±zα/2p^(1−p^)n\hat{p} \pm z_{\alpha/2} \sqrt{\frac{\hat{p}(1-\hat{p})}{n}}。
  2. 樣本大小的決定: 為了達到特定的誤差範圍(margin of error, MOE),樣本大小 nn 的計算公式為 n=(zα/2MOE)2p^(1−p^)n = \left(\frac{z_{\alpha/2}}{MOE}\right)^2 \hat{p}(1-\hat{p})。若無先驗的 p^\hat{p} 值,可使用 p^=0.5\hat{p}=0.5 以獲得最大樣本數。
  3. 比例的假設檢定: 比較兩個獨立樣本的比例,使用 zz-檢定。虛無假設 H0:p1=p2H_0: p_1 = p_2(或 p1≤p2p_1 \le p_2),對立假設 H1:p1>p2H_1: p_1 > p_2。檢定統計量為 z=p^1−p^2p^p(1−p^p)(1n1+1n2)z = \frac{\hat{p}_1 - \hat{p}_2}{\sqrt{\hat{p}_p(1-\hat{p}_p)(\frac{1}{n_1} + \frac{1}{n_2})}},其中 p^p=x1+x2n1+n2\hat{p}_p = \frac{x_1+x_2}{n_1+n_2} 為合併比例。

解題過程:

a) Find the 99% confidence interval for the current support of the bill.

目前(Present Sample)的數據:
樣本大小 n1=300n_1 = 300
支持人數 x1=150x_1 = 150
樣本比例 p^1=150300=0.5\hat{p}_1 = \frac{150}{300} = 0.5

要計算 99% 信賴區間,我們需要 zα/2z_{\alpha/2}。對於 99% 信賴水準,α=0.01\alpha = 0.01,α/2=0.005\alpha/2 = 0.005。查標準常態分配表,累積機率為 1−0.005=0.9951 - 0.005 = 0.995 時的 zz 值約為 2.5762.576。

信賴區間的公式:p^1±zα/2p^1(1−p^1)n1\hat{p}_1 \pm z_{\alpha/2} \sqrt{\frac{\hat{p}_1(1-\hat{p}_1)}{n_1}}
=0.5±2.5760.5(1−0.5)300= 0.5 \pm 2.576 \sqrt{\frac{0.5(1-0.5)}{300}}
=0.5±2.5760.25300= 0.5 \pm 2.576 \sqrt{\frac{0.25}{300}}
=0.5±2.5760.0008333= 0.5 \pm 2.576 \sqrt{0.0008333}
=0.5±2.576×0.0288675= 0.5 \pm 2.576 \times 0.0288675
=0.5±0.07447= 0.5 \pm 0.07447

信賴區間為 (0.5−0.07447,0.5+0.07447)=(0.42553,0.57447)(0.5 - 0.07447, 0.5 + 0.07447) = (0.42553, 0.57447)。
換成百分比,約為 (42.55%, 57.45%)。

b) Suppose the researcher wants to reduce the margin of error of the 95% confidence interval to be within ±2\pm 2 percentage points. How large a sample should be taken the next time when she conducts the survey?

目標是將 95% 信賴區間的誤差範圍(margin of error, MOE)控制在 ±0.02\pm 0.02 (2 percentage points)。
對於 95% 信賴水準,α=0.05\alpha = 0.05,α/2=0.025\alpha/2 = 0.025。查標準常態分配表,累積機率為 1−0.025=0.9751 - 0.025 = 0.975 時的 zz 值約為 1.961.96。

🔒

後續完整解題步驟與【答案】

免費註冊,享三天全站完整詳解閱覽。

免費註冊

第 3 題15 分

A salesperson contacts at most 10 potential customers every day. Each day, he randomly selects 10 individuals to be contacted and 15% of the contacted persons would make the purchase based on his past experience.

a) How many sales can he expect to make each day?
b) Suppose the manager promises him a bonus if he can make at least 3 sales in a day. What is the likelihood that the salesperson can earn the bonus?

登入後即可作答並保存紀錄。

這一題的完整詳解

本題主要考驗對二項分配的期望值與機率計算。

核心觀念:

  1. 二項分配: 若一隨機試驗包含 nn 次獨立的伯努利試驗,每次試驗成功的機率為 pp,則成功的次數 XX 服從二項分配 B(n,p)B(n, p)。
  2. 二項分配的期望值: E(X)=npE(X) = np。
  3. 二項分配的機率: P(X=k)=(nk)pk(1−p)n−kP(X=k) = \binom{n}{k} p^k (1-p)^{n-k}。

解題過程:

a) How many sales can he expect to make each day?

每天銷售員會隨機接觸 10 位潛在客戶。
這次接觸可以視為一次獨立的試驗。
每次接觸,客戶購買的機率是 p=0.15p = 0.15 (15%)。
接觸的客戶總數是 n=10n = 10。
銷售次數 XX 服從二項分配 B(n=10,p=0.15)B(n=10, p=0.15)。

銷售員每天預期能獲得的銷售次數(期望值)為:
E(X)=npE(X) = np
E(X)=10×0.15E(X) = 10 \times 0.15
E(X)=1.5E(X) = 1.5

所以,他每天預期可以獲得 1.5 次銷售。

b) Suppose the manager promises him a bonus if he can make at least 3 sales in a day. What is the likelihood that the salesperson can earn the bonus?

銷售員要獲得獎金,必須至少有 3 次銷售。這表示銷售次數 XX 需要滿足 X≥3X \ge 3。
我們需要計算 P(X≥3)P(X \ge 3)。
由於 XX 服從二項分配 B(10,0.15)B(10, 0.15),我們可以計算:
P(X≥3)=1−P(X<3)=1−[P(X=0)+P(X=1)+P(X=2)]P(X \ge 3) = 1 - P(X < 3) = 1 - [P(X=0) + P(X=1) + P(X=2)]

計算各項機率:
n=10n=10, p=0.15p=0.15, 1−p=0.851-p=0.85

🔒

後續完整解題步驟與【答案】

免費註冊,享三天全站完整詳解閱覽。

免費註冊

第 4 題15 分

A manager wants to compare the error rates of three auditors in the firm. Random samples of five performance reports were reviewed for each auditor. The results appear in the table below:

Error Rates for the Auditors

AuditorAuditor 1Auditor 2Auditor 3
1249
1583
1365
1457
1744

a) Set up the complete ANOVA Table.
b) Test at the 5% significance level to determine whether the mean error rates for the three auditors differ.

登入後即可作答並保存紀錄。

這一題的完整詳解

本題考驗單因子變異數分析 (One-Way ANOVA) 的設定與應用。

核心觀念:

  1. 單因子變異數分析 (One-Way ANOVA): 用於比較三個或三個以上獨立樣本的平均數是否存在顯著差異。
  2. ANOVA 的假設:
    • 各組樣本來自常態分配的母體。
    • 各組母體變異數相等(同質性)。
    • 各組樣本是獨立的。
  3. ANOVA 的邏輯: 將總變異分解為組間變異 (Between-group variation) 和組內變異 (Within-group variation)。若組間變異相對於組內變異足夠大,則可推論各組平均數有顯著差異。
  4. F 檢定統計量: F=MSBetweenMSWithinF = \frac{MS_{Between}}{MS_{Within}}。
  5. ANOVA 表: 總結變異分析結果,包含變異來源 (Source of Variation)、平方和 (Sum of Squares, SS)、自由度 (Degrees of Freedom, df)、均方 (Mean Square, MS)、F 值及 P 值。

解題過程:

首先,我們需要整理數據並計算一些基本統計量。
共有 k=3k=3 組(三位審計員)。
每組樣本數 ni=5n_i = 5。
總樣本數 N=k×ni=3×5=15N = k \times n_i = 3 \times 5 = 15。

各組數據:
Auditor 1 (X1X_1): 12, 15, 13, 14, 17
Auditor 2 (X2X_2): 4, 8, 6, 5, 4
Auditor 3 (X3X_3): 9, 3, 5, 7, 4

計算各組均值:
Xˉ1=12+15+13+14+175=715=14.2\bar{X}_1 = \frac{12+15+13+14+17}{5} = \frac{71}{5} = 14.2
Xˉ2=4+8+6+5+45=275=5.4\bar{X}_2 = \frac{4+8+6+5+4}{5} = \frac{27}{5} = 5.4
Xˉ3=9+3+5+7+45=285=5.6\bar{X}_3 = \frac{9+3+5+7+4}{5} = \frac{28}{5} = 5.6

計算總均值:
XˉTotal=∑Xˉik=14.2+5.4+5.63=25.23=8.4\bar{X}_{Total} = \frac{\sum \bar{X}_i}{k} = \frac{14.2 + 5.4 + 5.6}{3} = \frac{25.2}{3} = 8.4
或者 XˉTotal=∑XijN=71+27+2815=12615=8.4\bar{X}_{Total} = \frac{\sum X_{ij}}{N} = \frac{71+27+28}{15} = \frac{126}{15} = 8.4

a) Set up the complete ANOVA Table.

計算平方和 (Sum of Squares, SS):

  1. 組間平方和 (SSB, Sum of Squares Between groups):
    SSB=∑i=1kni(Xˉi−XˉTotal)2SSB = \sum_{i=1}^k n_i (\bar{X}_i - \bar{X}_{Total})^2
    SSB=5×(14.2−8.4)2+5×(5.4−8.4)2+5×(5.6−8.4)2SSB = 5 \times (14.2 - 8.4)^2 + 5 \times (5.4 - 8.4)^2 + 5 \times (5.6 - 8.4)^2
    SSB=5×(5.8)2+5×(−3.0)2+5×(−2.8)2SSB = 5 \times (5.8)^2 + 5 \times (-3.0)^2 + 5 \times (-2.8)^2
    SSB=5×33.64+5×9.00+5×7.84SSB = 5 \times 33.64 + 5 \times 9.00 + 5 \times 7.84
    SSB=168.2+45.0+39.2=252.4SSB = 168.2 + 45.0 + 39.2 = 252.4

  2. 組內平方和 (SSW, Sum of Squares Within groups):
    SSW=∑i=1k∑j=1ni(Xij−Xˉi)2SSW = \sum_{i=1}^k \sum_{j=1}^{n_i} (X_{ij} - \bar{X}_i)^2
    對於 Auditor 1:
    (12−14.2)2+(15−14.2)2+(13−14.2)2+(14−14.2)2+(17−14.2)2(12-14.2)^2 + (15-14.2)^2 + (13-14.2)^2 + (14-14.2)^2 + (17-14.2)^2
    =(−2.2)2+(0.8)2+(−1.2)2+(−0.2)2+(2.8)2= (-2.2)^2 + (0.8)^2 + (-1.2)^2 + (-0.2)^2 + (2.8)^2
    =4.84+0.64+1.44+0.04+7.84=14.8= 4.84 + 0.64 + 1.44 + 0.04 + 7.84 = 14.8

    對於 Auditor 2:
    (4−5.4)2+(8−5.4)2+(6−5.4)2+(5−5.4)2+(4−5.4)2(4-5.4)^2 + (8-5.4)^2 + (6-5.4)^2 + (5-5.4)^2 + (4-5.4)^2
    =(−1.4)2+(2.6)2+(0.6)2+(−0.4)2+(−1.4)2= (-1.4)^2 + (2.6)^2 + (0.6)^2 + (-0.4)^2 + (-1.4)^2
    =1.96+6.76+0.36+0.16+1.96=11.2= 1.96 + 6.76 + 0.36 + 0.16 + 1.96 = 11.2

    對於 Auditor 3:
    (9−5.6)2+(3−5.6)2+(5−5.6)2+(7−5.6)2+(4−5.6)2(9-5.6)^2 + (3-5.6)^2 + (5-5.6)^2 + (7-5.6)^2 + (4-5.6)^2
    =(3.4)2+(−2.6)2+(−0.6)2+(1.4)2+(−1.6)2= (3.4)^2 + (-2.6)^2 + (-0.6)^2 + (1.4)^2 + (-1.6)^2

🔒

後續完整解題步驟與【答案】

免費註冊,享三天全站完整詳解閱覽。

免費註冊

第 5 題35 分

To predict the stopping distance of their automobiles, a car company collects a sample of 50 observations and develops a simple linear regression model to test the relationship between a vehicle's speed and its stopping distance. The variable "stopping distance" (denoted by distance) is measured in feet and the variable "speed" (denoted by speed) is measured in miles per hour (mph). After the estimation, the software package produced the following output:

EstimateStandard Error
Intercept-17.57916.7584
Speed3.93240.4155

Residual standard error: 15.38
Average speed in the sample: 15.4
F-statistics: 89.567

Answer the following questions based on the given information.

a) What is the estimated regression equation?
b) What is the sample average of stopping distance?
c) At 0.05 level of significance, test whether the stopping distance is related to the vehicle's speed.
d) Provide the 95% confidence interval for the expected stopping distance when the driver is driving at 15.4 mph.
e) Compute the correlation coefficient between speed and distance.
f) Comment on the good of fit for the regression model.

登入後即可作答並保存紀錄。

這一題的完整詳解

核心觀念

本題考查**簡單線性迴歸分析(Simple Linear Regression, SLR)**的標準輸出報表解讀與推論,核心觀念包含:

  1. 迴歸估計式與迴歸直線性質:
    • 迴歸方程式型態:Y^i=β^0+β^1Xi\widehat{Y}_i = \hat{\beta}_0 + \hat{\beta}_1 X_i。
    • 最小平方法(OLS)性質:迴歸直線必通過樣本平均數點 (Xˉ,Yˉ)(\bar{X}, \bar{Y}),即 Yˉ=β^0+β^1Xˉ\bar{Y} = \hat{\beta}_0 + \hat{\beta}_1 \bar{X}。
  2. 斜率之假設檢定:
    • 檢定兩變數間是否有線性關係:H0:β1=0H_0: \beta_1 = 0 vs. H1:β1≠0H_1: \beta_1 \neq 0。
    • 檢定統計量可用 tt 檢定(t=β^1−0SE(β^1)t = \frac{\hat{\beta}_1 - 0}{SE(\hat{\beta}_1)})或 ANOVA FF 檢定(在簡單線性迴歸下,F=t2F = t^2),自由度為 n−2n - 2。
  3. 條件期望值之信賴區間(Confidence Interval for Mean Response):
    • 當 X=x0X = x_0 時,平均反應值 E(Y∣X=x0)E(Y|X = x_0) 的估計式為 μ^Y∣x0=β^0+β^1x0\hat{\mu}_{Y|x_0} = \hat{\beta}_0 + \hat{\beta}_1 x_0。
    • 其估計量之標準誤為: SE(μ^Y∣x0)=se1n+(x0−Xˉ)2∑(Xi−Xˉ)2SE(\hat{\mu}_{Y|x_0}) = s_e \sqrt{\frac{1}{n} + \frac{(x_0 - \bar{X})^2}{\sum (X_i - \bar{X})^2}}
    • 當 x0=Xˉx_0 = \bar{X} 時,第二項為 00,標準誤簡化為 SE(μ^Y∣Xˉ)=senSE(\hat{\mu}_{Y|\bar{X}}) = \frac{s_e}{\sqrt{n}}。
  4. 相關係數與判定係數:
    • 判定係數 R2=SSRSSTR^2 = \frac{SSR}{SST}。在簡單線性迴歸中,模型整體檢定的 FF 統計量滿足: F=R2/1(1−R2)/(n−2)  ⟹  R2=FF+(n−2)F = \frac{R^2 / 1}{(1 - R^2) / (n - 2)} \implies R^2 = \frac{F}{F + (n - 2)}
    • 樣本相關係數 rXYr_{XY} 與判定係數的關係為 rXY=sgn(β^1)R2r_{XY} = \text{sgn}(\hat{\beta}_1) \sqrt{R^2}。
  5. 模型配適度評估(Goodness of Fit):
    • 以判定係數 R2R^2、殘差標準誤 ses_e、以及整體顯著性 FF 檢定綜合評判。

解題方法與步驟推導

(a) 估計之迴歸方程式

  • 由報表可得:
    • 截距項估計值:β^0=−17.5791\hat{\beta}_0 = -17.5791
    • 斜率估計值:β^1=3.9324\hat{\beta}_1 = 3.9324
  • 依題目定義,應變數為 distance\text{distance},自變數為 speed\text{speed},估計之迴歸方程式為: distance^=−17.5791+3.9324×speed\widehat{\text{distance}} = -17.5791 + 3.9324 \times \text{speed}

(b) 煞車距離之樣本平均數 Yˉ\bar{Y}

  • 由 OLS 的基本幾何性質,估計迴歸線必通過重心 (Xˉ,Yˉ)(\bar{X}, \bar{Y}): Yˉ=β^0+β^1Xˉ\bar{Y} = \hat{\beta}_0 + \hat{\beta}_1 \bar{X}
  • 題目已知平均車速 Xˉ=15.4 mph\bar{X} = 15.4\text{ mph},代入計算: Yˉ=−17.5791+3.9324×15.4=−17.5791+60.55896=42.97986≈42.98 (feet)\bar{Y} = -17.5791 + 3.9324 \times 15.4 = -17.5791 + 60.55896 = 42.97986 \approx 42.98\text{ (feet)}

(c) 在 α=0.05\alpha = 0.05 下,檢定煞車距離是否與車速相關

  1. 建立假說: H0:β1=0vs.H1:β1≠0H_0: \beta_1 = 0 \quad \text{vs.} \quad H_1: \beta_1 \neq 0
  2. 檢定統計量:
    • 方法一(tt 檢定): t∗=β^1−0SE(β^1)=3.93240.4155≈9.464t^* = \frac{\hat{\beta}_1 - 0}{SE(\hat{\beta}_1)} = \frac{3.9324}{0.4155} \approx 9.464
    • 方法二(FF 檢定):
      報表直接給定 F∗=89.567F^* = 89.567。注意:(t∗)2=(9.464)2≈89.57(t^*)^2 = (9.464)^2 \approx 89.57,兩者完全等價。
  3. 拒絕域與臨界值:
    • 樣本數 n=50n = 50,自由度 df=n−2=50−2=48df = n - 2 = 50 - 2 = 48。
    • 顯著水準 α=0.05\alpha = 0.05。
    • 雙尾檢定臨界值 t0.025,48≈2.011t_{0.025, 48} \approx 2.011(或利用標準常態臨界值 z0.025≈1.96z_{0.025} \approx 1.96 近似;若用 FF 檢定,F0.05(1,48)≈4.04F_{0.05}(1, 48) \approx 4.04)。
  4. 決策與結論:
    • 因為 ∣t∗∣=9.464>2.011|t^*| = 9.464 > 2.011(或 F∗=89.567>4.04F^* = 89.567 > 4.04),p-value<0.001p\text{-value} < 0.001。
    • 故在 α=0.05\alpha = 0.05 顯著水準下拒絕 H0H_0。
    • 結論:有充分的統計證據顯示,車輛的煞車距離與行駛車速呈顯著的線性相關。

(d) 當車速為 15.4 mph 時,平均煞車距離的 95% 信賴區間

  1. 點估計:
    • 指定車速 x0=15.4 mphx_0 = 15.4\text{ mph},剛好等於樣本平均車速 Xˉ=15.4\bar{X} = 15.4。
    • 點估計值為: μ^Y∣x0=15.4=Yˉ=42.98 (feet)\hat{\mu}_{Y|x_0 = 15.4} = \bar{Y} = 42.98\text{ (feet)}
  2. 估計量標準誤:
    • 平均反應值之估計標準誤公式為: SE(μ^Y∣x0)=se1n+(x0−Xˉ)2∑(Xi−Xˉ)2SE(\hat{\mu}_{Y|x_0}) = s_e \sqrt{\frac{1}{n} + \frac{(x_0 - \bar{X})^2}{\sum (X_i - \bar{X})^2}}
    • 由於 x0=Xˉ=15.4x_0 = \bar{X} = 15.4,分子 (x0−Xˉ)2=0(x_0 - \bar{X})^2 = 0,公式大幅簡化為:
🔒

後續完整解題步驟與【答案】

免費註冊,享三天全站完整詳解閱覽。

免費註冊

其他考古題

112 年政治大學的其他科目

政治大學《統計學》其他年度