Multi-class text classification in NLP - Cheat Sheet & Quick Revision

Recall & Review

beginner

What is multi-class text classification?

It is a task where a text is sorted into one of many possible categories. For example, sorting emails into 'work', 'personal', or 'spam'.

intermediate

Why do we use softmax activation in multi-class classification models?

Softmax turns model outputs into probabilities that add up to 1, helping the model pick the most likely class.

beginner

Name a common loss function used for multi-class text classification.

Cross-entropy loss is commonly used because it measures how close the predicted probabilities are to the true class labels.

beginner

How does tokenization help in text classification?

Tokenization breaks text into smaller pieces like words or subwords, making it easier for the model to understand and learn from the text.

intermediate

What metric would you use to evaluate a multi-class text classification model?

Accuracy is common, but also precision, recall, and F1-score per class help understand model performance better.

What does multi-class text classification predict?

A Only two categories

B Multiple labels at once

C One label from multiple possible categories

D No labels, just text generation

Multi-class classification assigns one label from many possible categories to each text.

Which activation function is typically used in the output layer for multi-class classification?

A ReLU

B Tanh

C Sigmoid

D Softmax

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

What is the purpose of cross-entropy loss in multi-class classification?

A Measures distance between words

B Measures difference between predicted and true class probabilities

C Calculates accuracy

D Normalizes input text

Cross-entropy loss measures how close predicted probabilities are to the actual class labels.

Which step comes first in preparing text for classification?

A Tokenization

B Evaluation

C Model training

D Prediction

Tokenization breaks text into pieces so the model can process it.

Which metric gives a balanced view of precision and recall in multi-class classification?

A F1-score

B Softmax

C Loss

D Accuracy

F1-score balances precision and recall, useful for understanding model performance.

Explain the main steps involved in building a multi-class text classification model.

Think about how raw text becomes a prediction.

Describe why softmax activation and cross-entropy loss work well together in multi-class classification.

Focus on probability and loss relationship.