In: Statistics and Probability
Consumer price indices in 2018 in RK were:
Feb. 722.7
Mar. 726.5
Apr. 729.3
May 730.8
Jun. 732.2
Jul. 733.1
Aug. 734.2
Sep. 736.9
Oct. 739.9
Nov. 746.4
Dec. 751.5
a) Produce the simple linear regression analysis. Present the fitted line analytically. Is there a significant linear dependence of СPI on time? Explain.
b) Construct the 95% confidence predicted interval for the expected in the mean CPI for Jan. 2019 .
c) How much is r squared and what is its meaning?
d) What assumptions do you need for that prediction?
e) Draw the corresponding graph and show on it your predicted interval.
A)
x | y | (x-x̅)² | (y-ȳ)² | (x-x̅)(y-ȳ) |
1 | 722.7 | 25.00 | 147.95 | 60.82 |
2 | 726.5 | 16.00 | 69.95 | 33.45 |
3 | 729.3 | 9.00 | 30.95 | 16.69 |
4 | 730.8 | 4.00 | 16.51 | 8.13 |
5 | 732.2 | 1.00 | 7.09 | 2.66 |
6 | 733.1 | 0.00 | 3.11 | 0.00 |
7 | 734.2 | 1.00 | 0.44 | -0.66 |
8 | 736.9 | 4.00 | 4.15 | 4.07 |
9 | 739.9 | 9.00 | 25.36 | 15.11 |
10 | 746.4 | 16.00 | 133.09 | 46.15 |
11 | 751.5 | 25.00 | 276.77 | 83.18 |
ΣX | ΣY | Σ(x-x̅)² | Σ(y-ȳ)² | Σ(x-x̅)(y-ȳ) | |
total sum | 66 | 8083.5 | 110 | 715.4 | 269.60 |
mean | 6.00 | 734.86 | SSxx | SSyy | SSxy |
sample size , n = 11
here, x̅ = Σx / n= 6.00 ,
ȳ = Σy/n = 734.86
SSxx = Σ(x-x̅)² = 110.0000
SSxy= Σ(x-x̅)(y-ȳ) = 269.6
estimated slope , ß1 = SSxy/SSxx = 269.6
/ 110.000 = 2.4509
intercept, ß0 = y̅-ß1* x̄ =
720.1582
so, regression line is Ŷ =
720.1582 + 2.4509
*x
............
slope hypothesis test
tail=
2
Ho: ß1= 0
H1: ß1╪ 0
n= 11
alpha = 0.05
estimated std error of slope =Se(ß1) = Se/√Sxx =
2.464 /√ 110.00 =
0.2349
t stat = estimated slope/std error =ß1 /Se(ß1) =
2.4509 / 0.2349 =
10.4344
t-critical value= 2.262 [excel function:
=T.INV.2T(α,df) ]
Degree of freedom ,df = n-2= 9
p-value = 0.0000
decison : p-value<α , reject Ho
Conclusion: reject Ho and conclude that
linear relations exists between X and y
..............
B)
Predicted Y at X= 12
is
Ŷ = 720.15818 +
2.450909 * 12 =
749.569
For Individual Response Y
standard error, S(ŷ)=Se*√(1+1/n+(X-X̅)²/Sxx) =
2.9337
margin of error,E=t*std error=t*S(ŷ)=
2.2622 * 2.93 =
6.6366
Prediction Interval Lower Limit=Ŷ -E =
749.569 - 6.64 =
742.933
Prediction Interval Upper Limit=Ŷ +E =
749.569 + 6.64 =
756.206
............
c)
R² = (Sxy)²/(Sx.Sy) = 0.9236
there is 92.36% data is explained by time of CPI (y)
..........
d)
Linear relationship
Homoscedasticity
Multivariate normality
........
Please revert back in case of any doubt.
Please upvote. Thanks in advance.