You are on page 1of 6

Logit

Function
ODDS
The ratio of the probability of one event to that of an alternative
event. Ratio of success to failure.

Considering females group:

 we see that probability that a female will purchase


(success) the product is = 159/265

 Probability of failure (no purchase) for female is


106/265

 the odds is defined as (159/265)/(106/265) = 1.5.

 Higher the odds, better is the chance for success.

 Range of odds can be any number between [0 , ∞].


Log(ODDS)
It is a type of function that creates a map of probability
values from [0,1] to [- ∞ ,∞ ] .
It is also called as logit function.

If p=0,1-p=1:
If p=1,1-p=0
odds = 0/1
odds = 1/0
=0
= ∞
Log(odds) = log(0/1)
Log(odds) = log(1/0)
= log(0)-log(1)
= log(1)-log(0)
= ∞-0
= 0- ∞
=∞
=-∞
Logistic Regression and Logit function
 Thus the hypothesis in logistic regression will be :

h(X) = = P( y = 1 | X )

 This is called the logistic function or sigmoid function.


Logistic Regression and Logit function
 In logistic regression, the dependent variable is a logit, which is the natural log
of the odds.
 So if we are considering a binary classification problem (yes -1 or no -0 ) based
on a single independent variable X ,then :
 logit(P) = a + b X ----------(a)
 Our aim is to find the probability that the DV is 1 given a particular value for the
IV. This is done by applying a threshold on P.

 DV =

 So, how we will find P from (a) :


THANK YOU

You might also like