3 Classification Fundamentals Free Online FlashCards

What is classification?

Classification is supervised learning in which a model learns from labeled examples to predict categories for new examples.

What defines binary classification?

Binary classification predicts between two categories, such as spam and not spam.

What defines multiclass classification?

Multiclass classification predicts among three or more categories, such as cat, dog, or bird.

What are features in a classification example?

Features are the input information used to make a prediction, such as an email's text, sender, or number of links.

What is a label?

A label is the correct category that a model is meant to predict.

What is data leakage in classification?

Data leakage occurs when input information reveals the answer, making model evaluation misleading. Features should be available when the prediction is made.

What is training data used for?

Training data is used to fit the model and learn patterns connecting features to labels.

Why keep test data separate from training data?

Test data is held back to estimate performance on unseen examples. Testing on training examples can give overly optimistic results because a model may memorize them.

When using validation data, what is the test set reserved for?

Use a validation set to select settings or compare models, then reserve the test set for a final evaluation.

What do TP, FP, TN, and FN mean?

A true positive is a positive correctly predicted positive; a false positive is a negative incorrectly predicted positive. A true negative is a negative correctly predicted negative; a false negative is a positive incorrectly predicted negative.

How is classification accuracy calculated?

Accuracy is the fraction of all predictions that are correct: $\frac{TP + TN}{TP + TN + FP + FN}$.

What does precision measure?

Precision is the fraction of predicted positives that are truly positive: $\frac{TP}{TP + FP}$. High precision means fewer false alarms.