BUSINESS ANALYTICS WITH EXCEL FINAL EXAM ANSWER SHEET
NAME : joy maa kali
BATCH : MCA_3_CREDIT_BUSINESS_ANALYTICS_BATCH_FEB_12
COURSE : BUSINESS ANALYTICS WITH EXCEL
STREAM : MASTER OF COMPUTER APPLICATION
COLLEGE NAME : IITgoria ne mera dil tod diya
UNIVERSITY ROLL :
Question 1 : Performing Linear and Logistic regression on ICE CREAM DATA SET
Answer :
Both Linear and Logistic regression are done within the same excel sheet file for this question
First few entries in the dataset
Output Description :
On the top right we can that out of 500 samples our logistic regression predicted viability of 457 samples
correctly
Explanation :
Columns (for linear regression):
Temperature : already given in dataset . It is used as the independent variable
Revenue : already given in dataset, It is used as the dependent variable
Predicted Revenue : This is a custom made column, This is the dependent variable
whose value is calculated . The calculated revenue in this column is generated by using
the linear regression formula using the calculated slopes and intercepts from our
model .
Error : This is a custom made column, the calculated error for any row is the difference
between actual given revenue of that row and calculated revenue of that row
Squared Errors : The calculated value for any row is the square of that row’s errors
Columns (for logistic regression)
IsViable : This is a custom made column, for generating the discreet values to be used in
logistic regression.We create an assumption that an entry is economically viable only if
the actual given revenue of the entry is >= 500.
Probability(isViable) : Generated using the calculated revenue (value of the dependent
variable of that row) . To generate the probability we pass the estimated revenue into
the logistic function :
o P(x) = 1 / (1 + (e^(-y))) = (e^y) / (1 + (e^y))
IsViablePrediction : Generates the discreet value of economic viability of the entry ,
based on the calculated probability. We set a cutoff of 99 % probability for an entry to
be viable .
Match : The value of this column in 1 if actual viability and predicted viability matches ,
otherwise the value is 0
Output Description (formulas used) :
Correlation : CORRREL(independent variable range from data set , dependent variable
range from dataset)
Sum of squared error : SUM(squared error range)
Mean Squared Error : sum of Squared error cell ref / COUNT(range of data provided)
Root Mean Squared Error : SQRT(mean squared error cell reference)
Intercept : INTERCEPT(dependent variable range, independent variable range)
Slope : SLOPE(dependent variable range, independent variable range)
In the Scatter Plot the X-axis is the Temperature and the Y-axis is the Revenue
Question 2 : Performing Linear and Logistic regression on FUEL ECONOMY DATA
SET
Answer :
Both Linear and Logistic regression are done within the same excel sheet file for this question
First few entries in the dataset
Output Description :
On the top right we can that out of 100 samples our logistic regression predicted viability of 87 samples
correctly
Explanation :
Columns (for linear regression):
Hose Power : already given in dataset . It is used as the independent variable
Fuel Economy(MPG): already given in dataset, It is used as the dependent variable
Predicted MPG: This is a custom made column, This is the dependent variable whose
value is calculated . The calculated revenue in this column is generated by using the
linear regression formula using the calculated slopes and intercepts from our model .
ERROR: This is a custom made column, the calculated error for any row is the difference
between actual given revenue of that row and calculated revenue of that row
SQUARED ERRORS: The calculated value for any row is the square of that row’s errors
Columns (for logistic regression)
IsViable : This is a custom made column, for generating the discreet values to be used in
logistic regression. We create an assumption that an entry is economically viable only if
the actual given MPG of the entry is >= 20 MPG.
Probability(isViable) : Generated using the calculated revenue .To generate the
probability we pass the estimated MPG value into the logistic function :
o P(x) = 1 / (1 + (e^(-y))) = (e^y) / (1 + (e^y))
PredictedIsViable: Generates the discreet value of economic viability of the entry ,
based on the calculated probability. We set a cutoff of 90% of the calculated probability
for an entry to be viable .
Match : The value of this column in 1 if actual viability and predicted viability matches ,
otherwise the value is 0
Output Description (formulas used) :
Correlation : CORRREL(independent X range from data set , dependent Y range)
Slope : SLOPE(dependent variable range, independent variable range)
Intercept : INTERCEPT(dependent variable range, independent variable range)
Sum of squared error : SUM(squared error range)
Mean Squared Error : sum of Squared error cell ref / COUNT(range of data provided)
Root Mean Squared Error : SQRT(mean squared error cell reference)
In the Scatter Plot the X-axis is the Horse Power and the Y-axis is the Fuel Economy (MPG) , as
we can see that as the Horse Power increases the MPG value decreases.
PS: The solved excel files are attached in the mail body for reference.
