Inductive methods try to learn a general rule that can be reused on new data.

Inductive methods try to learn a general rule that can be reused on new data.

• A. Transductive semi-supervised learning
• B. Inductive semi-supervised learning
• C. Semi-supervised learning
• D. Supervised learning

Answer: (B)

Generalization is the central idea: learning a function from the training data that can predict labels for unseen inputs. In a semi-supervised setting, you have both labeled and unlabeled examples, and the goal is to leverage the unlabeled data to improve the learned function. When you hear “Inductive methods,” it signals that the model should generalize to new, unseen data, not just the instances from the current task. Transductive semi-supervised learning, by contrast, focuses on solving the specific test set you have at hand rather than producing a model that generalizes. So the combination of inductive and semi-supervised learning matches learning a reusable general rule from labeled and unlabeled data. Supervised learning uses labeled data only and also aims to generalize, but the statement emphasizes the inductive aspect with semi-supervised data, which is captured by inductive semi-supervised learning.

Generalization is the central idea: learning a function from the training data that can predict labels for unseen inputs. In a semi-supervised setting, you have both labeled and unlabeled examples, and the goal is to leverage the unlabeled data to improve the learned function. When you hear “Inductive methods,” it signals that the model should generalize to new, unseen data, not just the instances from the current task. Transductive semi-supervised learning, by contrast, focuses on solving the specific test set you have at hand rather than producing a model that generalizes. So the combination of inductive and semi-supervised learning matches learning a reusable general rule from labeled and unlabeled data. Supervised learning uses labeled data only and also aims to generalize, but the statement emphasizes the inductive aspect with semi-supervised data, which is captured by inductive semi-supervised learning.