109 年 國立清華大學計量財務金融研究所財務金融組《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

Suppose X1,X2,…,XnX_1, X_2, \ldots, X_n are drawn from independent and identical normal distribution N(μ,σ2)N(\mu, \sigma^2).
We define the sample mean as Xˉ=1n∑i=1nXi\bar{X} = \frac{1}{n}\sum_{i=1}^n X_i and the sample variance as s2=1n−1∑i=1n(Xi−Xˉ)2s^2 = \frac{1}{n-1}\sum_{i=1}^n (X_i - \bar{X})^2.

(a) Prove that E(s2)=σ2E(s^2) = \sigma^2. (10%)

(b) Define that W=∑i=1n(Xi−μ)(Xˉ−μ)σ2W = \frac{\sum_{i=1}^n (X_i - \mu)(\bar{X} - \mu)}{\sigma^2}. What is the distribution of W variable? (10%)

登入後即可作答並保存紀錄。

這一題的完整詳解

此題主要考察樣本變異數的期望值以及常態分布下相關變數的分布。

(a) 證明 E(s2)=σ2E(s^2) = \sigma^2
我們知道 s2=1n−1∑i=1n(Xi−Xˉ)2s^2 = \frac{1}{n-1}\sum_{i=1}^n (X_i - \bar{X})^2。
首先,我們將 ∑i=1n(Xi−Xˉ)2\sum_{i=1}^n (X_i - \bar{X})^2 展開:
∑i=1n(Xi−Xˉ)2=∑i=1n(Xi2−2XiXˉ+Xˉ2)\sum_{i=1}^n (X_i - \bar{X})^2 = \sum_{i=1}^n (X_i^2 - 2X_i\bar{X} + \bar{X}^2)
=∑i=1nXi2−2Xˉ∑i=1nXi+∑i=1nXˉ2= \sum_{i=1}^n X_i^2 - 2\bar{X}\sum_{i=1}^n X_i + \sum_{i=1}^n \bar{X}^2
=∑i=1nXi2−2Xˉ(nXˉ)+nXˉ2= \sum_{i=1}^n X_i^2 - 2\bar{X}(n\bar{X}) + n\bar{X}^2
=∑i=1nXi2−2nXˉ2+nXˉ2= \sum_{i=1}^n X_i^2 - 2n\bar{X}^2 + n\bar{X}^2
=∑i=1nXi2−nXˉ2= \sum_{i=1}^n X_i^2 - n\bar{X}^2

現在我們考慮期望值:
E(∑i=1n(Xi−Xˉ)2)=E(∑i=1nXi2−nXˉ2)E\left(\sum_{i=1}^n (X_i - \bar{X})^2\right) = E\left(\sum_{i=1}^n X_i^2 - n\bar{X}^2\right)
=E(∑i=1nXi2)−nE(Xˉ2)= E\left(\sum_{i=1}^n X_i^2\right) - nE(\bar{X}^2)

對於獨立且同分布的 Xi∼N(μ,σ2)X_i \sim N(\mu, \sigma^2),我們有 E(Xi)=μE(X_i) = \mu 和 Var(Xi)=E(Xi2)−(E(Xi))2=σ2Var(X_i) = E(X_i^2) - (E(X_i))^2 = \sigma^2。
因此,E(Xi2)=Var(Xi)+(E(Xi))2=σ2+μ2E(X_i^2) = Var(X_i) + (E(X_i))^2 = \sigma^2 + \mu^2。
所以,E(∑i=1nXi2)=∑i=1nE(Xi2)=∑i=1n(σ2+μ2)=n(σ2+μ2)=nσ2+nμ2E\left(\sum_{i=1}^n X_i^2\right) = \sum_{i=1}^n E(X_i^2) = \sum_{i=1}^n (\sigma^2 + \mu^2) = n(\sigma^2 + \mu^2) = n\sigma^2 + n\mu^2。

接著,我們計算 E(Xˉ2)E(\bar{X}^2)。我們知道 E(Xˉ)=E(1n∑i=1nXi)=1n∑i=1nE(Xi)=1n(nμ)=μE(\bar{X}) = E\left(\frac{1}{n}\sum_{i=1}^n X_i\right) = \frac{1}{n}\sum_{i=1}^n E(X_i) = \frac{1}{n}(n\mu) = \mu。
Var(Xˉ)=E(Xˉ2)−(E(Xˉ))2Var(\bar{X}) = E(\bar{X}^2) - (E(\bar{X}))^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 2 題

Suppose X1,X2,…,XnX_1, X_2, \ldots, X_n are drawn from independent and identical normal distribution N(μ,σ2)N(\mu, \sigma^2). There are two unbiased estimators for μ\mu:
μ^1=14(X1+0.5X2+1.5X3+X4)\hat{\mu}_1 = \frac{1}{4}(X_1 + 0.5X_2 + 1.5X_3 + X_4)
μ^2=14(0.5X1+X2+X3+1.5X4)\hat{\mu}_2 = \frac{1}{4}(0.5X_1 + X_2 + X_3 + 1.5X_4)
Compute the correlation coefficient between μ^1\hat{\mu}_1 and μ^2\hat{\mu}_2. (20%)

登入後即可作答並保存紀錄。

這一題的完整詳解

此題考驗對線性組合的期望值、變異數以及共變異數的理解,進而計算相關係數。

首先,我們確認 μ^1\hat{\mu}_1 和 μ^2\hat{\mu}_2 是 μ\mu 的不偏估計量。
對於 μ^1\hat{\mu}_1:
E(μ^1)=E(14(X1+0.5X2+1.5X3+X4))E(\hat{\mu}_1) = E\left(\frac{1}{4}(X_1 + 0.5X_2 + 1.5X_3 + X_4)\right)
=14(E(X1)+0.5E(X2)+1.5E(X3)+E(X4))= \frac{1}{4}(E(X_1) + 0.5E(X_2) + 1.5E(X_3) + E(X_4))
=14(μ+0.5μ+1.5μ+μ)= \frac{1}{4}(\mu + 0.5\mu + 1.5\mu + \mu)
=14(4μ)=μ= \frac{1}{4}(4\mu) = \mu。
μ^1\hat{\mu}_1 是不偏估計量。

對於 μ^2\hat{\mu}_2:
E(μ^2)=E(14(0.5X1+X2+X3+1.5X4))E(\hat{\mu}_2) = E\left(\frac{1}{4}(0.5X_1 + X_2 + X_3 + 1.5X_4)\right)
=14(0.5E(X1)+E(X2)+E(X3)+1.5E(X4))= \frac{1}{4}(0.5E(X_1) + E(X_2) + E(X_3) + 1.5E(X_4))
=14(0.5μ+μ+μ+1.5μ)= \frac{1}{4}(0.5\mu + \mu + \mu + 1.5\mu)
=14(4μ)=μ= \frac{1}{4}(4\mu) = \mu。
μ^2\hat{\mu}_2 是不偏估計量。

接下來,我們計算 μ^1\hat{\mu}_1 和 μ^2\hat{\mu}_2 的變異數。
由於 XiX_i 是獨立且同分布於 N(μ,σ2)N(\mu, \sigma^2),所以 Var(Xi)=σ2Var(X_i) = \sigma^2。
Var(μ^1)=Var(14(X1+0.5X2+1.5X3+X4))Var(\hat{\mu}_1) = Var\left(\frac{1}{4}(X_1 + 0.5X_2 + 1.5X_3 + X_4)\right)
=(14)2Var(X1+0.5X2+1.5X3+X4)= \left(\frac{1}{4}\right)^2 Var(X_1 + 0.5X_2 + 1.5X_3 + X_4)
=116(Var(X1)+(0.5)2Var(X2)+(1.5)2Var(X3)+Var(X4))= \frac{1}{16} (Var(X_1) + (0.5)^2 Var(X_2) + (1.5)^2 Var(X_3) + Var(X_4))
=116(σ2+0.25σ2+2.25σ2+σ2)= \frac{1}{16} (\sigma^2 + 0.25\sigma^2 + 2.25\sigma^2 + \sigma^2)
=116(1+0.25+2.25+1)σ2= \frac{1}{16} (1 + 0.25 + 2.25 + 1)\sigma^2
=4.516σ2=932σ2= \frac{4.5}{16}\sigma^2 = \frac{9}{32}\sigma^2。

Var(μ^2)=Var(14(0.5X1+X2+X3+1.5X4))Var(\hat{\mu}_2) = Var\left(\frac{1}{4}(0.5X_1 + X_2 + X_3 + 1.5X_4)\right)
=(14)2Var(0.5X1+X2+X3+1.5X4)= \left(\frac{1}{4}\right)^2 Var(0.5X_1 + X_2 + X_3 + 1.5X_4)
=116(Var(0.5X1)+Var(X2)+Var(X3)+Var(1.5X4))= \frac{1}{16} (Var(0.5X_1) + Var(X_2) + Var(X_3) + Var(1.5X_4))
=116((0.5)2Var(X1)+Var(X2)+Var(X3)+(1.5)2Var(X4))= \frac{1}{16} ((0.5)^2 Var(X_1) + Var(X_2) + Var(X_3) + (1.5)^2 Var(X_4))
=116(0.25σ2+σ2+σ2+2.25σ2)= \frac{1}{16} (0.25\sigma^2 + \sigma^2 + \sigma^2 + 2.25\sigma^2)
=116(0.25+1+1+2.25)σ2= \frac{1}{16} (0.25 + 1 + 1 + 2.25)\sigma^2
=4.516σ2=932σ2= \frac{4.5}{16}\sigma^2 = \frac{9}{32}\sigma^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 3 題

You plan to conduct a marketing experiment in which students are to taste one of two different brands of soft drink. Their task is to correctly identify the brand tasted. You select a random sample of 100 students and assume that the students have no ability to distinguish between the two brands. (Hint: If an individual has no ability to distinguish between the two soft drinks, then the two brands are equally likely to be selected.)

a. What is the probability that the sample will have between 50% and 60% of the identifications correct? (5%)
b. The probability is 90% that the sample percentage is contained within what symmetrical limits of the population percentage? (10%)
c. Which is more likely to occur-more than 65% correct identifications in the sample of 100 or more than 55% correct identifications in a sample of 500? Explain. (5%)

登入後即可作答並保存紀錄。

這一題的完整詳解

此題考驗對二項分布及其近似常態分布的應用,以及樣本比例的分布概念。

根據題意,學生無法分辨品牌,因此任意猜測一個品牌是正確的機率為 p=0.5p=0.5。這是一個二項分布問題。
令 nn 為樣本大小,XX 為樣本中猜對的學生數。則 X∼Binomial(n,p=0.5)X \sim Binomial(n, p=0.5)。
由於 nn 較大,我們可以使用常態分布來近似二項分布。
常態近似的條件是 np≥5np \ge 5 且 n(1−p)≥5n(1-p) \ge 5。
平均值為 E(X)=npE(X) = np,變異數為 Var(X)=np(1−p)Var(X) = np(1-p)。

a. 計算樣本中有 50% 到 60% 的辨識是正確的機率。
樣本大小 n=100n=100。p=0.5p=0.5。
np=100×0.5=50≥5np = 100 \times 0.5 = 50 \ge 5
n(1−p)=100×0.5=50≥5n(1-p) = 100 \times 0.5 = 50 \ge 5。
常態近似是適用的。
平均值 μ=np=100×0.5=50\mu = np = 100 \times 0.5 = 50。
變異數 σ2=np(1−p)=100×0.5×0.5=25\sigma^2 = np(1-p) = 100 \times 0.5 \times 0.5 = 25。
標準差 σ=25=5\sigma = \sqrt{25} = 5。

我們想計算 P(0.50≤Xn≤0.60)P(0.50 \le \frac{X}{n} \le 0.60),即 P(50≤X≤60)P(50 \le X \le 60)。
使用連續性修正 (continuity correction),我們將區間調整為 [50−0.5,60+0.5][50-0.5, 60+0.5],即 [49.5,60.5][49.5, 60.5]。
將 XX 轉換為標準常態變數 Z=X−μσZ = \frac{X - \mu}{\sigma}:
Z1=49.5−505=−0.55=−0.1Z_1 = \frac{49.5 - 50}{5} = \frac{-0.5}{5} = -0.1
Z2=60.5−505=10.55=2.1Z_2 = \frac{60.5 - 50}{5} = \frac{10.5}{5} = 2.1

所以,我們需要計算 P(−0.1≤Z≤2.1)P(-0.1 \le Z \le 2.1)。
P(−0.1≤Z≤2.1)=P(Z≤2.1)−P(Z≤−0.1)P(-0.1 \le Z \le 2.1) = P(Z \le 2.1) - P(Z \le -0.1)
查標準常態分布表:
P(Z≤2.1)≈0.9821P(Z \le 2.1) \approx 0.9821
P(Z≤−0.1)=P(Z≥0.1)=1−P(Z≤0.1)≈1−0.5398=0.4602P(Z \le -0.1) = P(Z \ge 0.1) = 1 - P(Z \le 0.1) \approx 1 - 0.5398 = 0.4602
P(−0.1≤Z≤2.1)≈0.9821−0.4602=0.5219P(-0.1 \le Z \le 2.1) \approx 0.9821 - 0.4602 = 0.5219。

【答案】約 0.5219

b. 在 90% 的機率下,樣本百分比落在人口百分比的哪個對稱區間內?
這裡我們考慮的是樣本比例 p^=X/n\hat{p} = X/n 的分布。
E(p^)=p=0.5E(\hat{p}) = p = 0.5。
Var(p^)=p(1−p)n=0.5×0.5100=0.25100=0.0025Var(\hat{p}) = \frac{p(1-p)}{n} = \frac{0.5 \times 0.5}{100} = \frac{0.25}{100} = 0.0025。
標準差 SD(p^)=0.0025=0.05SD(\hat{p}) = \sqrt{0.0025} = 0.05。
p^\hat{p} 近似服從 N(p,p(1−p)n)N(p, \frac{p(1-p)}{n})。

我們想找一個區間 [p−δ,p+δ][p - \delta, p + \delta],使得 P(p−δ≤p^≤p+δ)=0.90P(p - \delta \le \hat{p} \le p + \delta) = 0.90。

🔒

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

The per-store daily customer count (i.e., the mean number of customers in a store in one day) for a nationwide convenience store chain that operates nearly 10,000 stores has been steady, at 900, for some time. To increase the customer count, the chain is considering cutting prices for coffee beverages by approximately half. The small size will now be 0.59insteadof0.59 instead of 0.99, and the medium size will be 0.69insteadof0.69 instead of 1.19. Even with this reduction in price, the chain will have a 40% gross margin on coffee. To test the new initiative, the chain has reduced coffee prices in a sample of 36 stores, where customer counts have been running almost exactly at the national average of 900. After four weeks, the sample stores stabilize at a mean customer count of 964 and a standard deviation of 105. This increase seems like a substantial amount to you, but it also seems like a pretty small sample. Do you think reducing coffee prices is a good strategy for increasing the mean customer count?

a. State the null and alternative hypotheses and explain the meaning of the Type I and Type II errors in the context of this scenario. (5%)
b. At the 0.05 level of significance, is there evidence that reducing coffee prices is a good strategy for increasing the mean customer count? (5%)
c. Using normal distribution instead of student's t distribution, what is the probability of the Type II error if the true mean is 950. (5%)
d. How would it be possible to lower the chance of the Type I and Type II errors? (5%)

登入後即可作答並保存紀錄。

這一題的完整詳解

此題為一個應用題,考驗對假設檢定、Type I/II 錯誤的理解,以及常態分布與 t 分布在點估計和區間估計的應用。

根據題意,我們想檢定降低咖啡價格是否能增加平均顧客數。
原來的平均顧客數為 μ0=900\mu_0 = 900。
樣本大小 n=36n=36。
樣本平均數 xˉ=964\bar{x} = 964。
樣本標準差 s=105s = 105。
顯著水準 α=0.05\alpha = 0.05。

a. 虛無與對立假設,以及 Type I 和 Type II 錯誤的意義。
虛無假設 (H0H_0):降低咖啡價格不會增加平均顧客數(即平均顧客數維持不變或減少)。
對立假設 (H1H_1):降低咖啡價格會增加平均顧客數。

數學表示:
H0:μ≤900H_0: \mu \le 900
H1:μ>900H_1: \mu > 900
(或者,如果我們只關注是否「增加」,也可以寫成 H0:μ=900H_0: \mu = 900 vs H1:μ>900H_1: \mu > 900。但通常在檢定「增加」時,會將相等的情況也包含在虛無假設中,以求更保守的檢定。)

Type I 錯誤的意義:
在真實情況下,降低咖啡價格並不能增加平均顧客數(H0H_0 為真),但我們卻根據樣本數據得出結論,認為降低咖啡價格是個好策略(拒絕 H0H_0)。

Type II 錯誤的意義:
在真實情況下,降低咖啡價格確實能夠增加平均顧客數(H1H_1 為真),但我們卻根據樣本數據未能得出足夠證據證明這一點(未能拒絕 H0H_0)。

b. 在 α=0.05\alpha = 0.05 的顯著水準下,是否有證據顯示降低咖啡價格是個好策略?
由於樣本大小 n=36≥30n=36 \ge 30,且母體標準差未知,我們可以使用 t 檢定。
檢定統計量為 t=xˉ−μ0s/nt = \frac{\bar{x} - \mu_0}{s/\sqrt{n}}。
自由度為 df=n−1=36−1=35df = n-1 = 36-1 = 35。

計算檢定統計量:
t=964−900105/36=64105/6=6417.5=3.657t = \frac{964 - 900}{105/\sqrt{36}} = \frac{64}{105/6} = \frac{64}{17.5} = 3.657。

我們進行單尾檢定(右尾檢定),因為對立假設是 μ>900\mu > 900。
在 α=0.05\alpha = 0.05 和 df=35df=35 的情況下,查 t 分布表,臨界值 t0.05,35t_{0.05, 35} 約為 1.6901.690。
(如果查不到 35,可以用較保守的 30 或 40 的值,或者直接用常態近似值 1.645,但這裡有 t 值表,應該用 t 值。)
在提供的 t 分布表中,df=35 找不到,但 df=30 時 t0.05=1.697t_{0.05}=1.697,df=40 時 t0.05=1.684t_{0.05}=1.684。取 df=35 介於兩者之間,約為 1.690。

比較檢定統計量與臨界值:
t=3.657>1.690t = 3.657 > 1.690。
由於檢定統計量大於臨界值,我們拒絕虛無假設 H0H_0。

結論:在 0.05 的顯著水準下,有足夠的證據顯示降低咖啡價格確實能顯著增加平均顧客數。

【答案】
有證據顯示降低咖啡價格是個好策略。

c. 使用常態分布而非 t 分布,計算當真實平均數為 950 時,Type II 錯誤的機率。
這裡我們假設母體標準差 σ\sigma 等於樣本標準差 s=105s=105 (這是使用常態近似的常見做法,或者如果題目明確說母體已知 σ=105\sigma=105)。
我們現在要計算 Type II 錯誤的機率,即當真實平均數 μ=950\mu = 950 時,未能拒絕 H0:μ≤900H_0: \mu \le 900 的機率。

🔒

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

The owner of a restaurant serving Continental-style entrees has the business objective of learning more about the patterns of patron demand during the Friday-to-Sunday weekend time period. Data were collected from 630 customers on the type of entree and dessert ordered and organized into the following table:

TYPE OF ENTRÉE
TYPE OF
DESSERT Beef Poultry Fish Pasta Total
Ice cream 13 8 12 14 47
Cake 98 12 29 6 145
Fruit 8 10 6 2 26
None 124 98 149 41 412
Total 243 128 196 63 630

At the 0.05 level of significance, is there evidence of a relationship between type of dessert and type of entree? (20%)
(Critical Value: χ0.05,92=16.919,χ0.05,122=21.026,χ0.05,162=26.296\chi^2_{0.05, 9} = 16.919, \chi^2_{0.05, 12} = 21.026, \chi^2_{0.05, 16} = 26.296)

登入後即可作答並保存紀錄。

這一題的完整詳解

此題考驗對卡方獨立性檢定 (Chi-squared test of independence) 的應用。我們需要檢定兩種分類變數(甜點種類和主菜種類)之間是否存在關聯性。

檢定步驟如下:

  1. 建立假設:

    • 虛無假設 (H0H_0):甜點種類與主菜種類是獨立的(兩者之間沒有關聯性)。
    • 對立假設 (H1H_1):甜點種類與主菜種類不是獨立的(兩者之間存在關聯性)。
  2. 計算期望次數 (Expected Frequencies):
    在虛無假設成立的情況下,我們預期每一格的觀察次數應與其對應的邊際機率成比例。
    期望次數 Eij=(Row Total)i×(Column Total)jGrand TotalE_{ij} = \frac{(Row\ Total)_i \times (Column\ Total)_j}{Grand\ Total}。
    總樣本數 N=630N = 630。

    • Ice cream:
      • Beef: E11=47×243630≈18.10E_{11} = \frac{47 \times 243}{630} \approx 18.10
      • Poultry: E12=47×128630≈9.55E_{12} = \frac{47 \times 128}{630} \approx 9.55
      • Fish: E13=47×196630≈14.59E_{13} = \frac{47 \times 196}{630} \approx 14.59
      • Pasta: E14=47×63630=4.70E_{14} = \frac{47 \times 63}{630} = 4.70
    • Cake:
      • Beef: E21=145×243630≈55.87E_{21} = \frac{145 \times 243}{630} \approx 55.87
      • Poultry: E22=145×128630≈29.49E_{22} = \frac{145 \times 128}{630} \approx 29.49
      • Fish: E23=145×196630≈45.07E_{23} = \frac{145 \times 196}{630} \approx 45.07
      • Pasta: E24=145×63630=14.50E_{24} = \frac{145 \times 63}{630} = 14.50
    • Fruit:
      • Beef: E31=26×243630≈10.01E_{31} = \frac{26 \times 243}{630} \approx 10.01
      • Poultry: E32=26×128630≈5.28E_{32} = \frac{26 \times 128}{630} \approx 5.28
      • Fish: E33=26×196630≈8.08E_{33} = \frac{26 \times 196}{630} \approx 8.08
      • Pasta: E34=26×63630=2.60E_{34} = \frac{26 \times 63}{630} = 2.60
    • None:
      • Beef: E41=412×243630≈158.53E_{41} = \frac{412 \times 243}{630} \approx 158.53
      • Poultry: E42=412×128630≈83.97E_{42} = \frac{412 \times 128}{630} \approx 83.97
      • Fish: E43=412×196630≈128.34E_{43} = \frac{412 \times 196}{630} \approx 128.34
      • Pasta: E44=412×63630=41.20E_{44} = \frac{412 \times 63}{630} = 41.20
  3. 計算卡方檢定統計量:
    χ2=∑i,j(Oij−Eij)2Eij\chi^2 = \sum_{i,j} \frac{(O_{ij} - E_{ij})^2}{E_{ij}},其中 OijO_{ij} 是觀察次數,EijE_{ij} 是期望次數。

🔒

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