Microsoft Azure AI Fundamentals (AI-900)Describe fundamental principles of machine learning on AzureMedium
A financial institution wants to develop a machine learning model to detect fraudulent transactions. They have a large dataset of past transactions, with each transaction clearly labeled as 'fraudulent' or 'legitimate'. The primary goal is to accurately classify new, unseen transactions. Which type of machine learning is most appropriate for this scenario?
- AReinforcement Learning
- BSemi-Supervised Learning
- CSupervised Learning
- DUnsupervised Learning
Show answer & explanationAnswer & explanation
Correct answer: C. Supervised Learning
Since the dataset has clearly labeled data ('fraudulent' or 'legitimate'), the model can learn from these examples to classify new transactions. This is the definition of Supervised Learning.
Why the other options are wrong
- A. Reinforcement Learning involves an agent learning through trial and error with rewards, which is not suitable for this classification task.
- B. Semi-Supervised Learning uses both labeled and unlabeled data, but with a fully labeled dataset, Supervised Learning is more direct.
- D. Unsupervised Learning is used for unlabeled data to find hidden patterns, not for classification with known labels.
Supervised Learning
Supervised learning is a machine learning paradigm where an algorithm learns from a labeled dataset. It builds a model that maps input features to an output label, which can then be used to predict outcomes for new, unseen data.
- Requires labeled training data.
- Tasks include classification and regression.
- Aims to predict an output variable based on input variables.
- Learns from examples with known correct answers.
Memory trick: Machines learn in three main 'modes' — with a teacher, on their own, or by trying.