Scribd
Upload
19 views

Logistic Regression: Instructor: Karthik Sriram

Logistic regression is a statistical method used to model the probability of a binary response based on one or more predictor variables. It allows one to predict a discrete outcome, such as the likelihood of purchasing a product given certain variables like price. The method fits data to a logistic function using maximum likelihood estimation to determine the model's parameters. This better accounts for the binary nature of the response variable compared to linear regression, as the predicted probabilities remain between 0 and 1. The document provides an example problem demonstrating how logistic regression can be used to predict the probability of purchase at different price points.

Uploaded by

Anurag Rahul
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
19 views

Logistic Regression: Instructor: Karthik Sriram

Logistic regression is a statistical method used to model the probability of a binary response based on one or more predictor variables. It allows one to predict a discrete outcome, such as the likelihood of purchasing a product given certain variables like price. The method fits data to a logistic function using maximum likelihood estimation to determine the model's parameters. This better accounts for the binary nature of the response variable compared to linear regression, as the predicted probabilities remain between 0 and 1. The document provides an example problem demonstrating how logistic regression can be used to predict the probability of purchase at different price points.

Uploaded by

Anurag Rahul
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
You are on page 1/ 21

Logistic Regression

Instructor: Karthik Sriram

A simple example
A customers purchase decisions on a
product during her visits to a store are
shown below

A simple example-continued
Question: When the price is 2.5 how
likely is it that she will purchase ?

Trials/successes view of
data

Question: When the price is 2.5


how likely is it that she will
purchase?

Trials/successes view of
data

What may be the probability at


price=6 ?

We would like.
P(Purchase) = a function of
price
Can we answer this with linear
regression?

Fitting regression line

Some limitation of Linear


Regression
-Does not explicitly recognize 0-1
nature of the response
-Impact of change in price on
probability of purchase is different at
different levels of price. Linear
regression assumes it to be a constant.

A non-linear fit

Logistic Regression formulation


P(Purchase) = function of price
In Logistic Regression,

Logistic Regression formulation



Always between 0
and 1

Maximum Likelihood
Estimation
Compute the likelihood of observed
data

Maximize that likelihood to obtain


parameters

Likelihood for the example


data

Data

Find t.
e.g. P(purchase at price 2) =

For the example problem


-0.90
Predicted P(Purchase at price=2.5)
=

= =0.67

In practice
Need Sophisticated Numerical Methods
Minitab, R, SAS, SPSS etc.
Estimation of parameter is not enough,
Standard Errors also need to be
computed.

The MLE Idea in general


Least Squares Regression

Logistic Regression

Formulation

Formulation

takes values in ),

takes values in ,

(given the Xs)

(given the Xs)

Note that

Estimation: MLE same as Least


quares
Maximize

Estimation: MLE
Maximize:

Concepts
Formulation and Estimation
Interpretation of coefficients
Goodness of Fit
Validation

Practical Exercise
A direct marketing company..

Data snap shotdevelopment

Logistic Model Formulation


Logit

(Probability of buying)
=

Recall :

Parameter Estimates

You might also like

pFad - Phonifier reborn

Pfad - The Proxy pFad of © 2024 Garber Painting. All rights reserved.

Note: This service is not intended for secure transactions such as banking, social media, email, or purchasing. Use at your own risk. We assume no liability whatsoever for broken pages.


Alternative Proxies:

Alternative Proxy

pFad Proxy

pFad v3 Proxy

pFad v4 Proxy