1 Logistic Regression

This is binary classification

Logistic regression is a

Logistic regression predicts binary outcomes by modelling the probability using a

logistic function. (Classification mode)

Example of logistic regression

Spam or not spam

What is a Confusion Matrix in Multiclass Classification?

A confusion matrix for multiclass classification is a table that allows the visualization of the performance of a classification model by comparing predicted labels with actual labels across all classes.

For a multiclass problem with N classes, the matrix is an table where: Rows represent the true classes (actual labels). Columns represent the predicted classes (model's output). From this matrix, various metrics like accuracy, precision, recall, and F1 score can be calculated for each class individually

Sqaush line

Making the line limit the value beteween 0 & 1.

Logit Model

if op is between 0 & 1 (Logistic regrssion) then we call logit model. this happen when we attach sigmoid to the logistic regression.

Sigmoid function

This is help in calssifying the data with boundaries

Its range is 0 to 1.

Formula of sigmoid

How Logistic Regression Handles Classification Problems:

Logistic regression is a linear model that predicts probabilities for binary or multi-class classification problems. Instead of fitting a straight line, it uses the logistic (sigmoid) function to map the output of a linear equation to a probability range between 0 and 1.

Maximum Likelihood Estimation (MLE) vs Maximum a Posteriori Estimation (MAP):

MLE is purely data-driven, while MAP combines data with prior knowledge (e.g., from past observations or expert opinions).

what will be the output for each different z values

z = infintire then op is 1

z = 0 then op is 0.5

z = infinite then op is infinite

if we have 2 classes then its called

Binary classification

Ex : Dog 🐶 & Cat 🙀

when we have a lot of classes

Multiclass classification

multip label classification

we can see many label in same images

Type 1 vs Type 2

random state

This will give the distribution of data each time the same.

Hyperparameter tuning of the losgistic regression

• Grid search cv This will find the best parameters . But this take long time

• Random Search Cv - This will take less time as this just take sample of data.

How will you measure goodness of fit for logistic regression ?

Confusion Matrix and Classification Metrics

Softmax

multiclass classification algorithms used by logistic regression.

Algorithm/Context	Role of Softmax
Multinomial Logistic Regression	Converts logits into probabilities for multiple classes.
Neural Networks (e.g., CNNs, RNNs)	Used in the output layer for multiclass classification.
Natural Language Processing (NLP)	Converts model outputs into probabilities for text classification.