Multi-class classification in ML Python - Cheat Sheet & Quick Revision

Recall & Review

beginner

What is multi-class classification?

Multi-class classification is a type of machine learning task where the model must choose one label from three or more possible categories for each input.

beginner

Name a common algorithm used for multi-class classification.

Logistic regression with softmax, decision trees, random forests, and neural networks are common algorithms used for multi-class classification.

intermediate

What is the role of the softmax function in multi-class classification?

Softmax converts raw model outputs (logits) into probabilities that sum to 1, helping the model pick the most likely class.

intermediate

How do you evaluate a multi-class classification model?

You can use accuracy, confusion matrix, precision, recall, and F1-score for each class to evaluate performance.

beginner

Why is multi-class classification harder than binary classification?

Because the model must distinguish between more than two classes, making it more complex to learn and predict correctly.

Which function is commonly used to convert model outputs into probabilities in multi-class classification?

A ReLU

B Sigmoid

C Softmax

D Tanh

Softmax converts outputs into probabilities that sum to 1, suitable for multi-class tasks.

In multi-class classification, how many classes can the model predict?

A Only two classes

B Three or more classes

C Exactly three classes

D Only one class

Multi-class classification involves predicting one class out of three or more possible classes.

Which metric gives the overall percentage of correct predictions in multi-class classification?

A Accuracy

B Recall

C Precision

D Loss

Accuracy measures the percentage of correct predictions out of all predictions.

What does a confusion matrix show in multi-class classification?

A The dataset size

B Model training time

C The model's architecture

D How predictions match actual classes

A confusion matrix shows how many predictions were correct or wrong for each class.

Which of these is NOT a typical algorithm for multi-class classification?

A Linear Regression

B K-Nearest Neighbors

C Neural Networks

D Random Forest

Linear regression is for continuous outputs, not classification.

Explain how the softmax function helps in multi-class classification.

Think about turning raw scores into understandable chances.

Describe how you would evaluate a multi-class classification model's performance.

Consider metrics that show overall and class-specific results.