DECISION SCIENCE

Ques 1.

Part 1.
0.5 * 0.8 = 0.4
YES
0.8
0.5 * 0.2 = 0.1
X 0.2
NO
0.5 0.3 * 0.65 = 0.195
0.65 YES
0.3
Y 0.35 0.3 * 0.35 = 0.105
0.2 NO
0.2 * 0.4 = 0.08
0.4 YES
Z 0.6
NO 0.2 * 0.6 = 0.12

Part 2.

The following information, as interpreted from the probability tree diagram, is used to compute
the probability that it was airline A.

Airline A's scheduled flights are represented by Event A.

Airline B's scheduled flights are represented by Event B.

Airline C's scheduled flights are represented by the event C.

The event E denotes that the flight arrived on time.

The prior probabilities are:

P(A) = 0.50

P(B) = 0.30

P(C ) = 0.20

The posterior probabilities are:

P(E|A) = 0.80

P(E|B) = 0.65

P(E|C) = 0.40
We need to find the probability that it was Airline A which arrived on time. i.e., P(A|E).

Using Baye’s formula,

P(A|E) = P(E|A) x P(A)

P(E|A) x P(A) + P(E|B) x P(B) + P(E|C) x P(C)

= 0.80 * 0.50
0.80 * 0.50 + 0.65 *0.30 + 0.40 * 0.20

= 0.4
0.675

= 0.5925

Therefore, the probability that it was Airline A which left on time is 0.593.
Ques 2.

Part 1.

Independent Variables: Researchers alter or change independent variables, and the consequences
of these changes are recorded and compared. Predictor is another name for independent variables
(s). Independent variables are referred to as such because they forecast or anticipate the values of
the dependent variable in the model. These are denoted by ‘x’.

Dependent Variables: The term "dependent variable" refers to a sort of variable that assesses
how the independent variable(s) impact the test units. The dependent variables, on the other
hand, are the sorts of variables that are fully reliant on the independent variable (s). The The
dependent variable is also known as the predicted variable (s). The predictor / independent
variables predict or assume the values of the dependent variables, which are referred to as
dependent variables. These are denoted by ‘y’.

In the above question

Dependent Variable is Sales on Kahwa

Independent Variables are Number of Sales Representatives and Customer-Satisfaction Ratings

Part 2.

Regression analysis is a statistical approach that may be used to evaluate the hypothesis that one
or more variables are interdependent. In addition, regression analysis can assess the degree of the
effect of a change in one variable on another. Of course, this last trait is critical in projecting
future values. Regression analysis is based on the assumption that variables have a functional
connection and that the relationship is linear. This linearity assumption is necessary because, for
the most part, mathematicians and econometricians have yet to fully understand the theoretical
statistical features of non-linear estimation.

The simplest multiple regression model is one with two independent variables and one variable
with the maximum power (first-order regression model). The regression model is as follows:

y= β0 + β1 x1 + β2 x2 + €

Where,

y = the values of dependent variable i.e., Sales on Kahwa

β0 = the estimate of the regression constant

β1 = the partial regression coefficient for independent variable, x1

β2 = the partial regression coefficient for independent variable, x2

 x1 = Number of Sales Representatives

x2 = Customer-Satisfaction Ratings

€ = Error of prediction

The constant and coefficients are calculated using sample data, resulting in the model below.

yꞈ = b0 + b1 x1 + b2 x2

The simple regression equation for obtaining the sample slope and intercept is the outcome of
utilising calculus methods to reduce the regression model's sum of squares of error. Solving two
simultaneous equations with two unknowns b0 and b1 is required to construct these equations.
The values Σx , Σy, Σxy and Σyx are required to find the sample slope and intercept using these
formulae. Similarly, determining formulae to solve for multiple regression coefficients follows a
similar method. The formulae were created with the goal of reducing the sum of squares of error
for the model. As a result, the regression analysis presented here is known as least squares
analysis. For multiple regression studies with k independent variables, calculus methods are
used, resulting in k+1 equations with k+1 unknowns. As a result, a regression model with six
independent variables will yield seven simultaneous equations containing seven unknowns.

Simultaneous equations with three unknowns come from multiple regression models with two
independent variables (b0 , b1 and b2).

b0 n + b1 Σx1 + b2 Σx2 = Σy

b0 Σx1 + b1 Σx12+ b2 Σx1 x2 = Σx1y

b0 Σx2 + b1 Σx1 x2+ b2 Σx2 2= Σx2y

Solving these equations by hand is a difficult and time-consuming task. In a multiple regression
model with two independent variables, you must solve for the regression coefficients and
regression constant Σx1 , Σx2, Σx12 , Σx2 2, Σx1 x2, Σx1y and Σx2y. Almost all business researchers,
in fact, employ computer statistical software programmes to calculate regression coefficients,
regression constants, and other relevant data.

Part 3.

Summary Output

Regression Statistics
Multiple R 0.981088188
R Square 0.962534032
Adjusted R Square 0.943801048
Standard Error 3721.75587
Observations 10

ANOVA
Significance
  df SS MS F F
Regressio 213513900 51.3817787 0.00011341
n 3 2 711713000.8 8 4
83108800.5
Residual 6 5 13851466.76
221824780
Total 9 3      

Coeffi Standar Lower Upper Lower Upper

  cients d Error t Stat P-value 95% 95% 95.0% 95.0%
32355 8538.40 3.789 0.0090 11462. 53247. 11462. 53247.
Intercept .3 0022 382 79953 54873 97305 54873 97305
- -
Spending in advertise (in 1.686 1.50023 1.124 0.3038 1.9841 5.3577 1.9841 5.3577
INR) 84 2214 383 11184 0047 7149 00465 7149
- - - -
number of sales 155.5 1061.61 0.146 0.8883 2753.2 2442.1 2753.2 2442.1
representatives (person) 3 3404 503 22803 0391 44916 03914 44916
customer-satisfaction
ratings
(1=highly dissatisfied to 11309 2589.26 4.367 0.0047 4973.6 17645. 4973.6 17645.
5 = highly satisfied) .3 2959 781 30396 36202 03262 36202 03262

Part 4.

The regression equation is

yꞈ = 32355.3 + 1.68684x1 – 155.53x2 + 11309.3x3

Where,

yꞈ = prime interest rate

x1 = spending in advertise

x2 = number of sales representatives

x3 = customer-satisfaction ratings

The model indicates that for every one-unit (1%) increase in the spending in advertise, the
predicted prime rate interest rate increases by 32355.3. The model also indicates that for every
one-unit (1%) increase in number of sales representatives, the predicted prime interest decreases
by 155.53. The model also indicates that for every one-unit (1%) increase in customer-
satisfaction ratings, the predicted prime interest increases by 11309.3.
Ques 3(a).

75% Loves Reels , p= 0.75

25% Dislikes Reels , q= 0.25

Let X be the number of person loving reels.

Using Binomial Distribution

(a) P (X=15)
= 25 C15 * (0.75) 15 * (0.25)10

= 0.0416 (approx)

(b) P (X=20)

= 25
C20 * (0.75)20 * (0.25)5

= 0.16505 (approx)
Ques 3(b).

µ = 10,000

S.D. = 2400

As the number of hits follows Normal Distribution

(a) P (X > 12000)

P X- µ > 12000-10000
S.D. 2400

= 1 - P (Z < 0.83)

= 1 – 0.79673 (Using z score normal distribution table)

= 0.20327

(b) P (X < 9000)

P X-µ < 9000 - 10000

S.D. 2400

= 1 - P (Z < - 0.42)

= 1 – 0.66276 (Using z score normal distribution table)

= 0.33724