Cost-sensitive learning
Cost-sensitive learning incorporates unequal costs for different prediction errors into model training or decision-making.
Financial AI in Practice
Cost-sensitive learning incorporates differing costs of prediction errors into model training or decision-making.
- How it is used
- Addresses class imbalance in supervised financial fraud detection by making missed fraud more costly during learning.
- How deep it goes
- Defines unequal misclassification penalties and identifies a graph neural network loss modification, without specifying weights or implementation.
- Examples
- Illustrates the accuracy trap with a model that approves every transaction when at least 99% are legitimate.
- What you can do after
- Evaluate fraud-detection trade-offs using precision, recall, F1, and AUC-PR rather than accuracy alone.
Read how Financial AI in Practice explains cost-sensitive learning →
Build AI Drug Discovery Pipelines
Cost-sensitive learning incorporates unequal costs for different prediction errors into model training.
- How it is used
- Handles imbalanced inhibitor classification by making costly prediction errors influence model training.
- How deep it goes
- Explains cost matrices and class-balanced loss, but stops short of specifying a complete domain-specific cost matrix.
- Examples
- Uses virtual-screening inhibitor errors, a cost-matrix table, class-balanced formulas, and scikit-learn strategy comparisons.
- What you can do after
- Apply class weighting or explicit error costs and compare imbalance strategies with precision-recall area under the curve.
Read how Build AI Drug Discovery Pipelines explains cost-sensitive learning →
Related concepts
- class imbalanceaddresses · Class imbalance is a condition in a classification dataset where classes have substantially unequal numbers of examples.
- class-balanced losshas part · A class-balanced loss weights per-instance prediction errors to compensate for unequal class frequencies.
- cost matrixhas part · A cost matrix is a table whose entries specify the cost associated with combinations of alternatives, actions, assignments, or outcomes.
Questions about cost-sensitive learning
What is cost-sensitive learning?
Cost-sensitive learning incorporates unequal costs for different prediction errors into model training or decision-making.
Which Manning books explain cost-sensitive learning?
Financial AI in Practice and Build AI Drug Discovery Pipelines.
What is cost-sensitive learning a kind of?
Cost-sensitive learning is a kind of learning. Learning is the process by which a system acquires or updates knowledge, representations, or capabilities from information or outcomes in ways that can affect later behavior.
What are the parts of cost-sensitive learning?
Cost-sensitive learning includes class-balanced loss and cost matrix.
These guides are generated from the text of Manning books by Manning's knowledge graph and link to the sections they draw on. The full books are available on liveBook and at manning.com.

