Solution for Quantitative Methods for Decision Making Assignment

Group Number: 6

Group Members details:

RB20042 SACHIN S D

RB20043 Mohamed Basith A

RB20044 Harish M

RB20045 RAJESH P

RB20046 RENGASAMY B

RB20047 Ragunath Dhanabalan

RB20048 VINAY ANDRA

Question 1 (i):

Given:

Criteria Usage (in Yr.) Life (in Miles)

Early replacement buyers 0~3 <35000
Late replacement buyers >7 >73000

a) Confidence interval for the population mean no. of early replacement buyers

X N Σ Confidence interval α α/2 Z0.005

3.3 800 0.71 99% 0.01 0.005 2.575

Confiedence Interval=X ± ¿ ]

0.71
¿ 3.3 ±[ 2.575× ]
√ 800
¿ 3.3 ± 0.065

¿ [3.235 , 3.365]

b) Confidence interval for the population mean no. of Late replacement buyers

X N Σ Confidence interval Α α/2 Z0.005

4.3 500 0.66 99% 0.01 0.005 2.575

Confiedence Interval=X ± ¿ ]

0.66
¿ 4.3 ±[ 2.575 × ]
√500
¿ 4.3 ± 0.076

¿ [4.223 , 4.376]

C) Comparison of means with confidence interval

The confidence interval in both cases contains the population mean. So we are 99% confident
that the bayus classification of “early replacement buyers” & “late replacement buyers” with the given
criteria is valid as the population mean is in the confidence interval.
(ii):

Given:

Criteria In Sec.,
Mean alter time - Manual 15
With AI display panel <8

a) Population mean alert time for the new panel

X N S Confidence interval α α/2 t0.025 df

7.4 8 1.026 95% 0.05 0.025 2.145 n-1 = 14

Confiedence Interval=X ± ¿ ]

1.026
¿ 7.4 ±[ 2.145 × ]
√15
¿ 7.4 ± 0.568

¿ [6.832,7 .968]

b) The interval says the with AI panel, it is 95% confidence that the mean “alert time” is
between 6.832 to 7.968 seconds. Based on the interval the Ralph has strong evidence that the
new panel’s mean alter time is less than 8 seconds.
(iii)

Claim: National motor has equipped the ZX-900 to achieve shorter stopping distance than the
competitor (60ft)

a) Setting hypotheses

Null Hypotheses Alternative Hypotheses

H0 : µ ≥ 60 ft Ha : µ < 60 ft

b) Television network’s analysis of claim

X n σ α
57.8 81 6.02 0.05

X−µ
Z=
σ / √n

57.8−60
Z=
6.02/ √ 81
Z ¿−3.289 , Z0.05 = -1.645

P=0.00052 Which is less than the α = 0.05

So we can reject the null hypotheses in favour of alternative hypotheses and television network
has the strong evidence to reject the null hypotheses and will be advertising the claim.

(iv):

Claim: Firm’s new button’s mean lifetime exceeds the mean lifetime of the best remote button
currently available in market

a) Setting hypotheses

Null Hypotheses Alternative Hypotheses

H0 : µ ≤ 1200 Hrs Ha : µ > 1200 Hrs

X n Σ
1241.2 35 110.8

X−µ
Z=
σ / √n

1241.2−1200
Z=
110.8/ √ 35

Z=2.199

b) Critical value analysis using various α values

α - Value Z value Zα – Value Conclusion

0.1 2.199 1.282 Reject null hypotheses

0.05 2.199 1.645 Reject null hypotheses

0.01 2.199 2.326 Do not reject null hypotheses

0.001 2.199 3.090 Do not reject null hypotheses

c) When sample size changes

X n Σ
1241.2 100 110.8
1241.2−1200
Z=
110.8/ √ 100

Z=3.717

α - Value Z value Zα – Value Conclusion

0.1 3.717 1.282 Reject null hypotheses

0.05 3.717 1.645 Reject null hypotheses

0.01 3.717 2.326 Reject null hypotheses

0.001 3.717 3.090 Reject null hypotheses

It is evident that the increase in sample size n=100 increases the evidence of H a is true.

d) Practical Importance of results in (b) & (c)

Industry standard Claim % Increase Conclusion

1200Hrs 1241.2Hrs 3.43% Promotion of New button is depends on

the advertisement

So it is very difficult to promote the product to the customer by saying 3.4% improvement in
lifetime.

Also, results of (b) & (c) are important for the promotion of a product where television
networks use those results to confirm the claim.
In both sample sizes above 99% confidant interval we are rejecting the null hypothesis in
favour of alternate hypothesis i.e. increase mean life time which shows the same degree of
practical importance.

e) When change in sample mean & Standard deviation

X N Σ
1524.6 35 102.8

1524.6−1200
Z=
102.8/ √ 35

Z=18.68

Since our Z value is 18.08 which is greater than 4 , for values above 4 in the Z table the P value is
1.0000  . 
Since the alternate hypothesis is assumed for " > " mean lifetime value , the final P value will be 1 -
(1.0000) which is Zero . 
Clearly indicating a statistically insignificant result .  
But practically , The mean lifetime is more than 25% of the standard best life of button claimed earlier
which is very much visible to the customer .
Question 2

Answer for :: OxHome.csv problem

#R code for question 3

#Read OxHome.csv
Oxhome = read.csv(file.choose(), header=T)
attach(Oxhome)
names(Oxhome)

#Part 1 Simple Linear regression model for each variables

#Basic Scatter Plot for Sale vs sqft
plot(square.feet,sales.price, main = "Scatter plot of Sales vs. sqft")
Ans from R studio

#Fitting a linear model for sqft vs sales

sales.lm.1= lm(sales.price~square.feet)
summary(sales.lm.1)

Ans from R studio

lm(formula = sales.price ~ square.feet)

Residuals:
Min 1Q Median 3Q Max
-79.010 -17.411 0.281 12.838 85.577

Coefficients:
Estimate Std. Error t value Pr(>|t|)
(Intercept) -11.537932 14.929279 -0.773 0.443
square.feet 0.112158 0.009837 11.401 <2e-16 ***
---
Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1

Residual standard error: 32.33 on 58 degrees of freedom

Multiple R-squared: 0.6915, Adjusted R-squared: 0.6861
F-statistic: 130 on 1 and 58 DF, p-value: < 2.2e-16
Linear Regression Model equation for sale.price (Y) and square.feet (X1) is as
follows
Y = -11.538 + 0.112*X1
As T value is 11.401 and p value is 2.2 e-16 and Using H1:β1≠0, X1 is
significant Price and Square feet have association.

#Plotting the regression line for sqft vs sales

abline(sales.lm.1)
Ans from R studio

#Basic Scatter Plot for Sale vs rooms

plot(rooms,sales.price, main = "Scatter plot of Sales vs. rooms")
Ans from R studio

#Fitting a linear model for rooms vs sales

sales.lm.2= lm(sales.price~rooms)
summary(sales.lm.2)
Ans from R studio
lm(formula = sales.price ~ rooms)

Residuals:
Min 1Q Median 3Q Max
-89.621 -29.950 -6.719 19.660 136.940

Coefficients:
Estimate Std. Error t value Pr(>|t|)
(Intercept) 13.134 25.766 0.510 0.612
rooms 22.561 4.071 5.541 7.67e-07 ***
---
Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1

Residual standard error: 47.07 on 58 degrees of freedom

Multiple R-squared: 0.3461, Adjusted R-squared: 0.3349
F-statistic: 30.7 on 1 and 58 DF, p-value: 7.673e-07

Linear Regression Model equation for sale.price (Y) and rooms (X2) is as
follows
Y = 13.134 + 22.561*X2

As T value is 5.541 and p value is 7.67 e-7 and Using H1:β1≠0, X1 is

significant Price and rooms have association.

#Plotting the regression line for rooms vs sales

abline(sales.lm.2)
Ans from R studio

#Basic Scatter Plot for Sale vs bedrooms

plot(bedrooms,sales.price, main = "Scatter plot of Sales vs. bedrooms")
Ans from R studio
#Fitting a linear model for rooms vs sales
sales.lm.3= lm(sales.price~bedrooms)
summary(sales.lm.3)
Ans from R studio
lm(formula = sales.price ~ bedrooms)

Residuals:
Min 1Q Median 3Q Max
-74.693 -36.814 -8.048 20.952 151.952

Coefficients:
Estimate Std. Error t value Pr(>|t|)
(Intercept) 31.11 31.80 0.978 0.331932
bedrooms 41.65 10.72 3.885 0.000265 ***
---
Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1

Residual standard error: 51.85 on 58 degrees of freedom

Multiple R-squared: 0.2065, Adjusted R-squared: 0.1928
F-statistic: 15.09 on 1 and 58 DF, p-value: 0.000265

Linear Regression Model equation for sale.price (Y) and bedrooms (X3) is as
follows
Y = 31.11 + 41.65*X3
As T value is 11.401 and p value is 2.2 e-16 and Using H1:β1≠0, X1 is
significant Price and Square feet have association.

#Plotting the regression line for bedrooms vs sales

abline(sales.lm.3)
Ans from R studio
#Basic Scatter Plot for Sale vs age
plot(age,sales.price, main = "Scatter plot of Sales vs. age")
Ans from R studio

#Fitting a linear model for age vs sales

sales.lm.4= lm(sales.price~age)
summary(sales.lm.4)
Ans from R studio
lm(formula = sales.price ~ age)

Residuals:
Min 1Q Median 3Q Max
-73.97 -41.36 -14.96 22.94 154.31

Coefficients:
Estimate Std. Error t value Pr(>|t|)
(Intercept) 174.9518 12.1630 14.38 <2e-16 ***
age -0.7138 0.3038 -2.35 0.0222 *
---
Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1

Residual standard error: 55.62 on 58 degrees of freedom

Multiple R-squared: 0.08693, Adjusted R-squared: 0.07118
F-statistic: 5.522 on 1 and 58 DF, p-value: 0.02221

Linear Regression Model equation for sale.price (Y) and age (X4) is as follows
Y = 174.9518 - 0.7138*X4
As T value is -2.35 and p value is 0.022 and Using H0:β1=0, X1 is significant
Price and age has no association.

#Plotting the regression line for age vs sales

abline(sales.lm.4)
Ans from R studio

With considering following table

t
Variabl valu P
es e value RA2
11.4 2.00E- 0.686
Sqft 01 16 1
5.54 7.67E- 0.334
rooms 1 07 9
bedroo 3.88 0.0002 0.192
ms 5 65 8
- 0.071
age 2.35 0.0222 18

It is conclude that the model of sale price (Y) with Square feet (X1) Y = 13.134
+ 22.561*X2 is the best fir for considering single linear regression model.
Part 2:: Multiple regression model for sales data and it interpretation.
#Multiple regression model
sales.lm.mr= lm(sales.price~square.feet+rooms+bedrooms+age)
summary(sales.lm.mr)
Ans from R studio

lm(formula = sales.price ~ square.feet + rooms + bedrooms + age)

Residuals:
Min 1Q Median 3Q Max
-59.14 -14.68 -1.13 15.62 66.08

Coefficients:
Estimate Std. Error t value Pr(>|t|)
(Intercept) 19.48651 16.86523 1.155 0.25291
square.feet 0.10791 0.01228 8.788 4.62e-12 ***
rooms 11.88067 3.51702 3.378 0.00135 **
bedrooms -25.72796 8.13677 -3.162 0.00255 **
age -0.72074 0.15307 -4.709 1.73e-05 ***
---
Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1

Residual standard error: 26.21 on 55 degrees of freedom

Multiple R-squared: 0.8078, Adjusted R-squared: 0.7938
F-statistic: 57.77 on 4 and 55 DF, p-value: < 2.2e-16

Multiple Regression Model equation for sale.price (Y), square.feet(X1),

rooms(X2), Bedrooms(X3) and age (X4) is as follows
Y = 19.4865 + 0.107*X1 + 11.88067*X2 – 25.72796*X3 –
0.72074*X4
Part 3 : Stepwise Regression to find the best model
#step function to check AIC
step(sales.lm.mr)
Ans from R studio

Start: AIC=396.71
sales.price ~ square.feet + rooms + bedrooms + age

Df Sum of Sq RSS AIC

<none> 37781 396.71
- bedrooms 1 6868 44648 404.73
- rooms 1 7839 45619 406.02
- age 1 15230 53010 415.03
- square.feet 1 53049 90829 447.34

Call:
lm(formula = sales.price ~ square.feet + rooms + bedrooms + age)

Coefficients:
(Intercept) square.feet rooms bedrooms age
19.4865 0.1079 11.8807 -25.7280 -0.7207

Model equation for sale.price (Y), square.feet(X1), rooms(X2), Bedrooms(X3)

and age (X4) using AIC by step regression is as follows
Y = 19.4865 + 0.107*X1 + 11.88067*X2 – 25.72796*X3 –
0.72074*X4

From above details we conclude that 4 independent variables are significant on determining
the sales prices and it is the best fit model for the regression.

Part 4 : forecast price for square.feet(X1) 5000, rooms(X2) 20, Bedrooms(X3)

10 and age (X4) 100

Ans::
Y = 19.4865 + 0.107*X1 + 11.88067*X2 – 25.72796*X3 –0.72074*X4
Y=19.4865 + (0.107*5000) + (11.88067*20) – (25.72796*10) – (0.72074*100)

Sales Price for the house is as follows:: 462.7459

Since most of the independent variables are extrapolated and Sale price calculated is
considered as the bad forecast.

Question 3:
Part(a)(i) Model to Predict Sales Representative Performance using first
five independent variables:
#Attach Sales File

AData = read.csv(file.choose(), header=T)

attach(AData)
names(AData)

#Multiple regression model

Saleperform = lm(Sales~Time+MktPoten+Adver+MktShare+Change)
summary(Saleperform)

Coefficients:
Estimate Std. Error t value Pr(>|t|)
(Intercept) -1.114e+03 4.199e+02 -2.653 0.01571 *
Time 3.612e+00 1.182e+00 3.057 0.00649 **
MktPoten 4.209e-02 6.731e-03 6.253 5.27e-06 ***
Adver 1.289e-01 3.704e-02 3.479 0.00251 **
MktShare 2.570e+02 3.914e+01 6.566 2.76e-06 ***
Change 3.245e+02 1.573e+02 2.063 0.05301 .
---
Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1

Residual standard error: 430.2 on 19 degrees of freedom

Multiple R-squared: 0.915, Adjusted R-squared: 0.8926
F-statistic: 40.91 on 5 and 19 DF, p-value: 1.585e-09

Sales (Y)  Response

Time(X1), MktPoten(X2), Adver(X3), MktShare(X4), Change(X5)  Predictors

Estimated multiple regression model :

𝒀=−1114+3.612X1+0.04209X2+0.1289X3+257X4+324.5X5
Part(a)(ii) Model to Predict Sales Representative Performance using all
eight independent variables:
Saleperform2 =
lm(Sales~Time+MktPoten+Adver+MktShare+Change+Accts+WkLoad+Rating)
summary(Saleperform2)

Coefficients:
Estimate Std. Error t value Pr(>|t|)
(Intercept) -1.508e+03 7.786e+02 -1.936 0.070675 .
Time 2.010e+00 1.931e+00 1.041 0.313408
MktPoten 3.721e-02 8.202e-03 4.536 0.000338 ***
Adver 1.510e-01 4.711e-02 3.205 0.005518 **
MktShare 1.990e+02 6.703e+01 2.969 0.009041 **
Change 2.909e+02 1.868e+02 1.557 0.138983
Accts 5.551e+00 4.776e+00 1.162 0.262130
WkLoad 1.979e+01 3.368e+01 0.588 0.564896
Rating 8.189e+00 1.285e+02 0.064 0.949977
---
Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1

Residual standard error: 449 on 16 degrees of freedom

Multiple R-squared: 0.922, Adjusted R-squared: 0.8831
F-statistic: 23.65 on 8 and 16 DF, p-value: 1.816e-07

Sales (Y)  Response

Time(X1), MktPoten(X2), Adver(X3), MktShare(X4), Change(X5), Accts(X6), WkLoad(X7), Rating(X8)
 Predictors

Estimated multiple regression model:

𝒀=−1508+2.010X1+0.03721X2+0.151X3+199X4+290.9X5+5.5
51X6+19.79X7+8.189X8
Comparing both models model with 5 independent variants is better as it has Adjusted
R-squared = 0.8926 which is closer to 1 than with 8 predictors

Part(b)(i) Arriving at the best model using AIC:

#step function to check AIC

step(Saleperform2)

Start: AIC=312.2
Sales ~ Time + MktPoten + Adver + MktShare + Change + Accts +
WkLoad + Rating

Df Sum of Sq RSS AIC

- Rating 1 819 3226803 310.20
- WkLoad 1 69654 3295639 310.73
- Time 1 218443 3444428 311.83
<none> 3225985 312.20
- Accts 1 272416 3498400 312.22
- Change 1 488906 3714891 313.72
- MktShare 1 1777623 5003608 321.17
- Adver 1 2071255 5297240 322.60
- MktPoten 1 4148313 7374298 330.87

Step: AIC=310.2
Sales ~ Time + MktPoten + Adver + MktShare + Change + Accts +
WkLoad

Df Sum of Sq RSS AIC

- WkLoad 1 70565 3297369 308.74
- Time 1 229491 3456294 309.92
<none> 3226803 310.20
- Accts 1 289214 3516017 310.35
- Change 1 614867 3841671 312.56
- MktShare 1 1815755 5042559 319.36
- Adver 1 2343726 5570529 321.85
- MktPoten 1 4282223 7509026 329.32

Step: AIC=308.74
Sales ~ Time + MktPoten + Adver + MktShare + Change + Accts

Df Sum of Sq RSS AIC

- Accts 1 219521 3516890 308.36
<none> 3297369 308.74
- Time 1 326829 3624197 309.11
- Change 1 577624 3874992 310.78
- Adver 1 2459312 5756681 320.67
- MktShare 1 3515946 6813315 324.89
- MktPoten 1 4712574 8009943 328.93

Step: AIC=308.36
Sales ~ Time + MktPoten + Adver + MktShare + Change

Df Sum of Sq RSS AIC

<none> 3516890 308.36
- Change 1 788061 4304951 311.41
- Time 1 1729460 5246350 316.35
- Adver 1 2240625 5757515 318.68
- MktPoten 1 7236631 10753521 334.30
- MktShare 1 7979336 11496226 335.97

Call:
lm(formula = Sales ~ Time + MktPoten + Adver + MktShare + Change)

Coefficients:
(Intercept) Time MktPoten Adver MktShare Change
-1114 3.612 4.209e-02 1.289e-01 2.570e+02 3.245e+02

Best model:

𝒀=−1114+3.612X1+0.04209X2+0.1289X3+257X4+324.5X5
Part(b)(ii) Discussion on the significant predictors of the best model:
As per t test (at α= 1%) The p values obtained from t19 (refer RStudio Coeffecients for Best
Fitted Model) of MktShare(X4) is .00000276 and of MktPoten(X2) is .00000527 are < .01
and closer to 0 as compared to other variables.
Thus X4 and X2 are the most significant variables, which implies enough evidence of
association between Sales Performance and Market Share and Market Potential controlling
Time, Advertisement and Change