4 Machine Learning Basics Free Online FlashCards

What does machine learning use data to do?

Machine learning uses data to build models that make predictions or identify patterns.

What distinguishes supervised learning?

Supervised learning uses examples with input features and known target labels to learn predictions for new examples.

What kind of output does regression predict?

Regression predicts a numerical quantity, such as delivery time or house price.

What kind of output does classification predict?

Classification predicts a category, such as whether a transaction is fraudulent.

What does unsupervised learning find without target labels?

Unsupervised learning finds structure in data without target labels. A discovered cluster still needs interpretation in context.

What are model parameters, and how are they learned?

Parameters are learned from data; training adjusts them to reduce loss, a numerical measure of prediction error.

How do hyperparameters differ from learned parameters?

Hyperparameters are choices made outside the fitting process, such as learning rate or model complexity limits; they are typically tuned using validation data.

What roles do training, validation, and test data play?

Training data fits the model, validation data helps compare choices, and held-out test data provides a final performance estimate.

How should preprocessing be split to avoid data leakage?

Split data before learning preprocessing steps. Fit those steps on training data, then apply them to other splits to prevent information leakage.

How should time-dependent data be split for evaluation?

Evaluate on later observations rather than randomly mixing past and future, so the split reflects the time-dependent use case.

What is model generalization?

Generalization is the ability to make reliable predictions on new examples drawn from the conditions where the model will be used.

What is underfitting?

Underfitting occurs when a model fails to capture useful patterns and performs poorly even on training data.