Scribd
Upload
80 views

11-Logistic Regression

Logistic regression is a statistical method used for binary classification problems in machine learning. It models the probability that a data point belongs to one of two classes based on input features. Logistic regression can be applied when two classes can be divided by a straight line. Unlike linear regression, logistic regression's output is a probability between 0 and 1 using the sigmoid function. It also avoids issues like misclassification from outliers and probabilities outside of 0-1 that linear regression has.

Uploaded by

Naitik Tyagi
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
80 views

11-Logistic Regression

Logistic regression is a statistical method used for binary classification problems in machine learning. It models the probability that a data point belongs to one of two classes based on input features. Logistic regression can be applied when two classes can be divided by a straight line. Unlike linear regression, logistic regression's output is a probability between 0 and 1 using the sigmoid function. It also avoids issues like misclassification from outliers and probabilities outside of 0-1 that linear regression has.

Uploaded by

Naitik Tyagi
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/ 27

Logistic Regression

Logistic Regression
• Logistic regression is a statistical method used for binary classification
problems in machine learning.
• It works by modelling the probability that a given data point belongs to
one of two classes (e.g., 0 or 1, yes or no) based on the values of one or
more input features.
• Logistic regression algorithm can be applied to a problem statement
when two classes in the problem are linearly separable, i.e., they can be
divided by a straight line.
Why not Linear Regression
• Can we solve the classification problem using linear regression?
Why not Linear Regression
• Now, to solve this kind of problem using linear regression, our objective
is to draw the best-fit line so that the value of the loss function is
minimum
• Y=mx+c is the equation of the line which will divide the points into two
classes.
• If you set the condition that whenever the value of y >=.5, the person is
obese, otherwise the person is not obese.
Why not Linear Regression
• Now if a person with weight 60 comes then we find that y< 0.5, and we
will say that person is not obese.
• So, this means that with the help of linear regression we can classify
properly then question arises what is the need of logistic regression.
Why not Linear Regression
• To understand the need for logistic regression, we added one more
person to the dataset with a weight of 140.
Why not Linear Regression
• Now we have redrawn our new best-fit line due to the addition of a new
point.
• But this has resulted in misclassification as the person with weights like
75, 80, 85, and 100 are now on the left side and their corresponding
value of y is less than 0.5, and we declare them as not obese.
• So, this will result in a higher error rate.
• Also, the value of y can be greater than 1 and less than 0.
Why not Linear Regression
• So, the two important reasons why linear regression can’t be used for
classification problems are as follows:
• Whenever we have a lot of outlier points in the dataset best fit line will completely
deviate and it will give multiple incorrect results.
• Output value is probability but using linear regression, it can be greater than one
and less than zero.
• To solve the above 2 issues, we use logistic regression in place of linear
regression.
Sigmoid Function
• The sigmoid function maps the input to a value between 0 and 1. It is
given by: σ(x) = 1 / (1 + exp(-x))
• The sigmoid function was historically popular but is less common now
due to the vanishing gradient problem as it squashes the output.
• The sigmoid function is commonly used in the output layer of a binary
classification problem.
Logistic Regression-Hypothesis
Interpretation of Hypothesis Output
Decision Boundary
Decision Boundary: Linear
Decision Boundary: Non-Linear
Decision Boundary: Non-Linear
Cost Function
Cost Function
Cost Function
Cost Function
Cost Function
Gradient Descent
Gradient Descent
Multi-Class Classification: One vs All
Multi-Class Classification: One vs All
Multi-Class Classification: One vs All
Multi-Class Classification: One vs All
Multi-Class Classification: One vs All

You might also like