Scribd
Upload
47 views17 pages

11 - Logistic Regression

Logistic regression is a regression method used to predict the probability of a binary outcome. It models the probability of an event occurring as a function of independent variables. Some key aspects are: - It is used when the dependent variable is dichotomous (binary), such as pass/fail, yes/no. - It does not require a linear relationship between dependent and independent variables. - It estimates the odds ratio to predict the probability of an event based on maximum likelihood estimation. - The output is a predicted probability value between 0 and 1 rather than an integer like linear regression.

Uploaded by

Muhammad Waseem Anwar
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
47 views17 pages

11 - Logistic Regression

Logistic regression is a regression method used to predict the probability of a binary outcome. It models the probability of an event occurring as a function of independent variables. Some key aspects are: - It is used when the dependent variable is dichotomous (binary), such as pass/fail, yes/no. - It does not require a linear relationship between dependent and independent variables. - It estimates the odds ratio to predict the probability of an event based on maximum likelihood estimation. - The output is a predicted probability value between 0 and 1 rather than an integer like linear regression.

Uploaded by

Muhammad Waseem Anwar
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 17

LOGISTIC

REGRESSION

Sachin Nandakar
LOGISTIC REGRESSION
It is a regression method used to find
the probability of an event which is a
binary outcome.

In other words,
It is a regression analysis to conduct
when the dependent variable is
dichotomous(binary).
EXAMPLE

 How does the probability of placement


(Selected / Not-selected) change for every
additional percentage score in examination?

 Do change in body fat intake & age have an


influence on the probability of having a
heart attack (Yes / No)?
MAJOR ASSUMPTIONS &
UNDERSTANDINGS
 The dependent variable should be dichotomous in
nature (Ex: Pass/Fail, Present/Absent)
 There should be no high correlations (multi-
colinearity) among the predictors(independent
variables).
 It requires large sample size because maximum
likelihood estimates are less powerful at low sample
sizes.
 The regression doesn’t require linear relationship
between dependent and independent variables.
RANGE, ODDS & LOGIT
FUNCTION
Here the value of Y-dependent variable ranges
from 0 to 1 and it can be represented by following
equation
PLACEMENT DATA

Placement
Percentage
(0-Not Selected
Score
1-Selected)
42 0
48 1
54 1
65 1
40 0
52 0
60 1
58 1
40 0
SCATTER PLOT

BEST FIT ?
PROBABILITY
WHAT ARE THE ODDS?
ODDS RATIO
BRING BERNOULLI
PREDICTIVE EQUATION
After consideration of training datasets, we can
get the values bo & b1 –co-efficient of regression
& assuming the probability value is 0.76 for the
placement score of 55, then the outcome can be
approximated as 1 i.e selected.
LINEAR v/s LOGISTIC
Linear Regression Logistic Regression

Variable Type Continuous Dependent Categorical Dependent


Variable Variable

Estimation Least-Square Estimation Maximum Likelihood


Method Estimation

Equation y = ax + b y = b0 + b1x1 + b2x2 +….

Best fit line Straight Line Curve

Output Predicted Integer Value Predicted Binary Value (


0 or 1 )

You might also like