Chartered Financial Analyst (CFA) Practice Exam Level 2

What is an implication of using unsupervised learning?

1. The algorithm requires labeled training data

2. The algorithm learns patterns from unlabeled data

3. The algorithm cannot generalize from training data

4. The algorithm is always more effective than supervised learning

The correct answer is: The algorithm learns patterns from unlabeled data

Using unsupervised learning implies that the algorithm operates on unlabeled data to identify patterns and structures within the dataset. Unlike supervised learning, which relies on labeled inputs to predict outcomes, unsupervised learning enables the identification of hidden relationships or groupings without prior knowledge of the outputs. In unsupervised learning, techniques such as clustering and dimensionality reduction allow the algorithm to explore the data independently, leading to insights such as discovering natural groupings within the data or identifying anomalies. This exploratory approach is particularly useful in scenarios where labeling data is impractical or too costly. The other options are not accurate representations of unsupervised learning. The need for labeled data is a characteristic of supervised learning, and while it's true that unsupervised algorithms can struggle with generalization, this is not inherently related to unsupervised learning itself. Lastly, there's no basis for claiming that unsupervised learning is always more effective than supervised learning, as their effectiveness typically depends on the specific problem being addressed and the nature of the data available.