Decision Science

June 2021 Examination

1. Nemi Mehta is owning 50 Hectares of land near Junagadh. On his Farm, he cultivates Giloy,
which is a medicinal plant in Ayurveda. The below-given table shows the sales of ‘Giloy Vati’
that his team is preparing as per one Ayurveda script (book), and the amount they spend on the
Advertisement of it.

Nemi’s problem is to analyze the effect of Advertisement on sales. Firstly, He wants to

understand the presence of a linear relationship between the sales and ‘amount spent in
advertisement’. He also wants to run a correlation and regression to know whether he should
keep spending money on Advertisements or not. If sales figures are not affected by
advertisement, he should not spend money on it.

So, take the EXCEL output of the following and write the interpretation.

 Appropriate type of Graph which can exhibit the linear relation b/w variables.

 Karl Pearson’s Correlation Coefficient

 Regression Analysis

 Graph that shows predicted v/s actual sales figures.

Advertising

Region code Sales (TV spots per

1 260.3 month) 5(in INR,
2 286.1 7
3 279.4 6
4 410.8 9
5 438.2 12
6 315.3 8
7 565.1 11
8 570.0 16
9 426.1 13
10 315.0 7
11 403.6 10
12 220.5 4
13 343.6 9
14 644.6 17
15 520.4 19
 16 329.5 9
17 426.0 11
18 343.2 8
19 450.4 13
20 421.8 14

Answer 1
The first study is talking about the details of the graphical presentation of the sales with the
advertisement.

The purpose of the study is also that we want to check that is there any linear relationship exist
between these two.

 Appropriate Graph type which shows the linear relation b/w variables is as follows

The following is the sales details of the various regions.

Sales ( INR 000s)

700

600

500

400

300

200

100

0
0 5 10 15 20 25

The following is the detail of the advertising details is as follows:

 Advertising ( TV spots per month) ( in INR, 000s)
20

18

16

14

12

10

0
0 5 10 15 20 25

We can see that the relationship is linear between the Advertising cost as well as the sales which has
been achieved.

The linear relation graph is as follows between the sales and the advertising cost is as follows:

Both the sales as well as the advertising cost have been plotted as follows:

Linear relation graph

700

600

500

400

300

200

100

0
0 5 10 15 20 25

Sales ( INR 000s)

We have plotted the graph of the sales as well as the advertisement and we have seen that the graph is
moving in the same directions. This shows that there isa linear relationship is there between sales and
advertisement. Our second study is about the correlation coefficient between these two variables.
The correlation coefficient will tell us the relationship between these data. Is there are positive
correlation exists or is there a negative correlation exists or is there is no correlation exists between
them.

 Karl Pearson correlation coefficient is as follows :

The formula for the Karl person coefficient of the correlation is as follows:

r = n ∑ XY −¿ ¿ ¿
Region Sales ( INR Advertising ( TV spots per month)
Code 000s) ( in INR, 000s)
x y xx xy yy
1 260.3 5 67756.09 1301.5 25
2 286.1 7 81853.21 2002.7 49
3 279.4 6 78064.36 1676.4 36
4 410.8 9 168756.6 3697.2 81
5 438.2 12 192019.2 5258.4 144
6 315.3 8 99414.09 2522.4 64
7 565.1 11 319338 6216.1 121
8 570 16 324900 9120 256
9 426.1 13 181561.2 5539.3 169
10 315 7 99225 2205 49
11 403.6 10 162893 4036 100
12 220.5 4 48620.25 882 16
13 343.6 9 118061 3092.4 81
14 644.6 17 415509.2 10958.2 289
15 520.4 19 270816.2 9887.6 361
16 329.5 9 108570.3 2965.5 81
17 426 11 181476 4686 121
18 343.2 8 117786.2 2745.6 64
19 450.4 13 202860.2 5855.2 169
20 421.8 14 177915.2 5905.2 196
Sum 7969.9 208 3417395 90552.7 2472

∑X = 7969.90

∑y = 208

∑XX = 3717395

∑XY = 90552.7

∑YY = 2472
 r = n ∑ XY −¿ ¿ ¿
20 X 90552.7−7969.90 X 208
r= √¿¿¿

1811054−1657739.20
= √ ¿ ¿¿

153314.80 153314.80
= √¿¿¿
= 258606.64 = 0.5928

We have calculated the correlation coefficient between these two variables and these two
variables are positive and it is 0.5928.

This means we can say that there isa positive correlation and having a direct impact. This means as
we increase the advertisement cost the sales increase and as we decrease the advertisement cost the
sales decreases. Our third study is to find out the regression equation of X on the Y.

This means we want to have the regression equation of the sales based on the advertisement.

 Regression Analysis of sales on Advertising

Regression analysis of X on Y is as follows

( X-X bar ) = by ( Y – Year )

by = n ∑ XY −¿ ¿ ¿
20 X 90552.7−7969.90 X 208 153314.80
bxy = 20 X 2472− ( 208¿ ¿¿ 2 ) = 49440−( 43264 )

= 153314.80 / 6176 = 24.82

Average ox X = 398.495

Average oy Y = 10.40

( X – 398.495 ) = 24.82 ( Y – 10.40 )

X – 398.495 = 24.82 Y – 258.17

X = 24.82 Y + 398.495 – 258.17

X = 24.82 Y + 140.325

X = 140.325 + 24.82 Y
Our third study of the regression equation has been calculated and the regression equation has been
found.

The excel data has been as follows for the prediction.

Region Code Sales ( INR 000s) Trend = X = 140.325 + 24.82 Y

1 260.3 264.43
2 286.1 314.07
3 279.4 289.25
4 410.8 363.71
5 438.2 438.17
6 315.3 338.89
7 565.1 413.35
8 570 537.45
9 426.1 462.99
10 315 314.07
11 403.6 388.53
12 220.5 239.61
13 343.6 363.71
14 644.6 562.27
15 520.4 611.91
16 329.5 363.71
17 426 413.35
18 343.2 338.89
19 450.4 462.99
20 421.8 487.81

 Following is the graph for the predicted V/s

Predicted V/S Actual Sales figures

700

600

500

400

300

200

100

0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

Sales ( INR 000s) Trend = X = 140.325 + 24.82 Y

2. The table given below is the ‘single year age population’ (taken from census 2011). This table
shows the population of people (age-wise) living in Leh at the time of the census 2011 survey.
Transform this ungrouped data into Grouped data by forming age groups (only five age groups),
and then find out Mean, Variance, Standard deviation, Ogive, and Histogram. Write the
summary based on your calculations.
Age Age Age Age Age
in Populati in Populati in Populati in Populati in Populati
Yea on Yea on Yea on Yea on Yea on
rs rs rs rs rs
11 1998 22 2839 33 2696 44 1652
1 1958 12 1916 23 2935 34 2781 45 1806
2 1725 13 2138 24 3601 35 2799 46 1460
3 1814 14 2139 25 4110 36 2450 47 1226
4 1768 15 2096 26 4089 37 2142 48 1225
5 1871 16 2044 27 3716 38 2114 49 1006
6 1888 17 2027 28 3702 39 1725 50 1454
7 1768 18 2065 29 3084 40 2218
8 1712 19 2013 30 3475 41 1802
9 1780 20 2459 31 2844 42 1751
10 1862 21 2594 32 2684 43 1659

Source: Data.Gov.in

Answer 2.
The table given below is as follows for the single year age population that has been taken from census
2011, the table which has been taken age-wise is as follows.

Age in Populat Age in Populat Age in Populat Age in Populat Age in Populat
years ions years ions years ions years ions years ions
1 1958 11 1998 21 2594 31 2844 41 1802
2 1725 12 1916 22 2839 32 2684 42 1751
3 1814 13 2138 23 2935 33 2696 43 1659
4 1768 14 2139 24 3601 34 2781 44 1652
5 1871 15 2096 25 4110 35 2799 45 1806
6 1888 16 2044 26 4089 36 2450 46 1460
7 1768 17 2027 27 3716 37 2142 47 1226
8 1712 18 2065 28 3702 38 2114 48 1225
9 1780 19 2013 29 3084 39 1725 49 1006
10 1862 20 2459 30 3475 40 2218 50 1454
18146 20895 34145 24453 15041

For the mean calculation, all the data has been summarized as below,

After summarization of the mean data, the mean has been calculated and the mean is as follows :

class interval age group Mid Value ( X) Population (f) X*F

1 to 10 5 18146 90730
11 to 20 15 20895 313425
21 t0 30 25 34145 853625
31 to 40 35 24453 855855
41 to 50 45 15041 676845
Sum 112680 2790480

Mean value =
∑ fx = 2790480 / 112680 = 24.76
∑F

Variance

The data has been summarized for the standard variation and the standard deviation has been
calculated as follows :

Mid Value Population X - ( X -Xbar ) ( X- Xbar

class interval age group ( X) (f) X*F Xbar )
1 to 10 5 18146 90730 -19.76 390.46
11 to 20 15 20895 313425 -9.76 95.26
21 t0 30 25 34145 853625 0.24 0.06
31 to 40 35 24453 855855 10.24 104.86
41 to 50 45 15041 676845 20.24 409.66
279048
Sum 50 112680 0 1000.29

∑ ( x−xbar )2 = 1000.29
Standard deviation =
√ n √ 112680
= 0.094

Ogive graph is as follows:

The data for the Less than Ogive has been summarized and the graph has been prepared which is as
follows:

class interval age group Mid Value ( X) Population (f) Less than Cum. Frequency
1 to 10 5 18146 18146
11 to 20 15 20895 39041
21 t0 30 25 34145 73186
31 to 40 35 24453 97639
41 to 50 45 15041 112680
Sum 50 112680

It is less than Ogive graph is as follows:

Less than
Ogive graph

120000

100000

80000

60000

40000

20000

0
1 to 10 11 to 20 21 t0 30 31 to 40 41 to 50

Data for the more than ogive graph is as follows

class interval age group Mid Value ( X) Population (f) More than Cum. Frequency
1 to 10 5 18146 112680
11 to 20 15 20895 94534
21 t0 30 25 34145 73639
31 to 40 35 24453 39494
41 to 50 45 15041 15041
Sum 50 112680

Following is the more than Ogive Graph is as follows:

More than Ogive graph

120000

100000

80000

60000

40000

20000

0
1 to 10 11 to 20 21 t0 30 31 to 40 41 to 50

The data for the Histogram is as follows:

class interval age group Mid Value ( X) Population (f)

1 to 10 5 18146
11 to 20 15 20895
21 t0 30 25 34145
31 to 40 35 24453
41 to 50 45 15041

The Histogram graph is as follows:

 Population (f)
40000
35000
30000
Population (f)

25000
20000
15000
10000
5000
0
1 to 10 11 to 20 21 t0 30 31 to 40 41 to 50
class interval age group

Conclusion:

We have calculated the mean and found the mean age for the details provided is as follows,

The mean age is 24.76

And the standard deviation is 0.094. This shows that when the data for the 112680 number has been
collected the mean age is 24.76 between the age of 1 year to 50 years.

The deviation which has been calculated and found to be very less and is 0.094.

Hence we can say that the deviation is quite less.

Hence we data are averagely divided between all the age groups and this could be seen from the
deviation also.

The graph of the histogram is also average compare to the age of 21 to 30 years where the frequency of
the data is maximum. Rest all the data are under the average figure only.

3.a. The given table shows the rainfall of Gujarat Region. Forecast the rainfall using Exponential
Smoothing. Use Alpha =0.2, 0.5 and 0.8. Data is available from 1997 to 2016, use this series for
the calculation and forecast the rainfall for the year 2017. To know, what extent the prediction is
correct, actual rainfall for 2017 (1024.4 millimeters) is provided. Based on MSE and MAD, find
out which alpha values among the three suggestions are relatively near to actual value?
SUBDIVISION YEAR ANNUAL (in MM)
 Gujarat Region 1997 1068.9

Gujarat Region 1998 1070

Gujarat Region 1999 568.4

Gujarat Region 2000 550.6

Gujarat Region 2001 849

Gujarat Region 2002 637.2

Gujarat Region 2003 1160.3

Gujarat Region 2004 1005.8

Gujarat Region 2005 1316.4

Gujarat Region 2006 1478

Gujarat Region 2007 1178.9

Gujarat Region 2008 911.1

Gujarat Region 2009 641.6

Gujarat Region 2010 1088.7

Gujarat Region 2011 890.5

Gujarat Region 2012 714

Gujarat Region 2013 1118.6

Gujarat Region 2014 705.7

Gujarat Region 2015 622.9

Gujarat Region 2016 764.9

Source: Data.Gov.in

Answer 3 a:

Calculation of the forecast when the alpha value is 0.20, 0.50, and 0.80 is as follows:

By using the formula of the exponential smoothing is as follows:

Ft = α x A t-1) + ( 1- α ) X ( F( t- 1) )

Where the alpha value is 0.20, 0.50 as well as the 0.80

Forecast when Forecast when Forecast when

Subdivision Year Annual ( in MM) α = 0.20 α = 0.50 α = 0.80
Gujrat Region 1997 1068.9 1068.9 1068.9 1068.9
Gujrat Region 1998 1070 1068.90 1068.90 1068.90
Gujrat Region 1999 568.4 1069.12 1068.90 1068.90
Gujrat Region 2000 550.6 968.98 1069.01 1068.90
Gujrat Region 2001 849 885.30 1018.99 1068.99
Gujrat Region 2002 637.2 878.04 952.15 1028.99
Gujrat Region 2003 1160.3 829.87 915.09 967.52
Gujrat Region 2004 1005.8 895.96 872.48 925.58
Gujrat Region 2005 1316.4 917.93 884.22 883.10
Gujrat Region 2006 1478 997.62 901.07 884.00
Gujrat Region 2007 1178.9 1093.70 949.35 897.66
Gujrat Region 2008 911.1 1110.74 1021.52 939.01
Gujrat Region 2009 641.6 1070.81 1066.13 1005.02
Gujrat Region 2010 1088.7 984.97 1068.47 1053.91
Gujrat Region 2011 890.5 1005.71 1026.72 1065.56
Gujrat Region 2012 714 982.67 1016.22 1034.49
Gujrat Region 2013 1118.6 928.94 999.44 1019.87
Gujrat Region 2014 705.7 966.87 964.19 1003.53
Gujrat Region 2015 622.9 914.64 965.53 972.06
Gujrat Region 2016 764.9 856.29 940.08 966.84
Gujrat Region 2017 838.01 898.19 945.43

The calculation of the MSE, as well as MAD, is as follows when the alpha = 0.20

Calculation of the MAD when the alpha is 0.20

Forecast
when
Subdivisio Annual Error ( At - Absolute( At ( At- Ft) ( At
n Year ( in MM) α = 0.20 Ft ) - Ft ) - ft )
Gujrat
Region 1997 1068.9 1068.9 0.00 0.00 0.00
Gujrat
Region 1998 1070 1068.90 1.10 1.10 1.21
Gujrat
Region 1999 568.4 1069.12 -500.72 500.72 250720.52
Gujrat
Region 2000 550.6 968.98 -418.38 418.38 175038.48
Gujrat
Region 2001 849 885.30 -36.30 36.30 1317.75
Gujrat
Region 2002 637.2 878.04 -240.84 240.84 58004.21
Gujrat 2003 1160.3 829.87 330.43 330.43 109182.32
Region
Gujrat
Region 2004 1005.8 895.96 109.84 109.84 12065.26
Gujrat
Region 2005 1316.4 917.93 398.47 398.47 158781.20
Gujrat
Region 2006 1478 997.62 480.38 480.38 230763.86
Gujrat
Region 2007 1178.9 1093.70 85.20 85.20 7259.57
Gujrat
Region 2008 911.1 1110.74 -199.64 199.64 39855.14
Gujrat
Region 2009 641.6 1070.81 -429.21 429.21 184221.24
Gujrat
Region 2010 1088.7 984.97 103.73 103.73 10760.33
Gujrat
Region 2011 890.5 1005.71 -115.21 115.21 13274.36
Gujrat
Region 2012 714 982.67 -268.67 268.67 72184.39
Gujrat
Region 2013 1118.6 928.94 189.66 189.66 35971.97
Gujrat
Region 2014 705.7 966.87 -261.17 261.17 68209.65
Gujrat
Region 2015 622.9 914.64 -291.74 291.74 85109.79
Gujrat
Region 2016 764.9 856.29 -91.39 91.39 8351.89
Gujrat
Region 2017 1024.4 838.01 186.39 186.39 34740.89
Sum -968.06 4738.47 1555814.03

The MAD =
∑ absolute (At −Ft ) = 4738.47 / 21 = 225.64
n

The MSE =
∑ ( At −Ft )( At −Ft ) = 1555814.03 / 21 = 74086.38
n

The calculation of the MAD and the MSE when the alpha is 0.50 and 0.80 is as follows :

Calculation of the MAD when the alpha is 0.50

Forecast
when
Subdivisio Yea Annual ( in Error ( At - Ft Absolute( At - ( At- Ft) ( At -
n r MM) α = 0.50 ) Ft ) ft )
Sum -1369.66 5137.42 1845305.24

Calculation of the MAD when the alpha is 0.80

Sum -1571.24 5161.51 1919112.25

The MAD when alpha is 0.50 = 5137.42 / 21 = 244.64

The MSE when the alpha is 0.50= 1845305.24 / 21= 87871.68

The MAD when the alpha is 0.80 = 5161.51 / 21 = 245.79

The MSE when the alpha is 0.80 = 1919112.25 / 21 = 91386.30

The alpha value of 0.20 is the relatively near to the actual values.

3.b. A new gas-electric hybrid car has recently hit the market. The distance traveled on 1 gallon
of fuel is normally distributed with a mean of 65 miles and a standard deviation of 4 miles. Find
the probability of the following events. (show the concerned region by z curve)

1. The car travels more than 70 miles per gallon.

2. The car travels less than 60 miles per gallon.

3. The car travels between 55 and 70 miles per gallon.

(5 Marks)

Answer 3 b:

The distance traveled on 1-gallon of fuel is normally distributed with a mean of 65 miles and a
standard deviation of 4 miles.

The Z curve when the car travels more than 70 Miles per gallon.

The Z value for the car that travels more than 70 miles per gallon is as follows:

¯¿ 70−65
Z = X −X Standard deviation ¿ = 4
= 1.25

The value of the 1.25 in the Z table is = (0.8461 + 0.8485) / 2 = 0.8473

 More than 70
Normal Curve

70

Mean 65

(2) The car travels less than 60 miles per gallon

Normal curve Z curve

Less than the 60 60

65 mean
value

3:
The car travels between 55 and 70 miles per gallon is as follows:

More than
55 and less
Normal Z curve than 70

55 65 70