Local explanation
A local explanation describes how or why a model produced a particular prediction or decision for an individual input.
Ensemble Methods for Machine Learning
Local explainability, or local interpretability, is the ability to explain why a model produced a particular prediction for an individual input.
- How it is used
- Supports trust, action, debugging, failure analysis, and bias discovery by accounting for individual predictions.
- How deep it goes
- Contrasts local and global scope, then details LIME surrogates, SHAP attribution, and direct EBM component contributions.
- Examples
- Uses diabetes diagnosis, a four-person prize coalition, and Bank Marketing test example 3104.
- What you can do after
- Apply it via LIME, SHAP, or InterpretML to inspect feature and interaction contributions for an individual prediction.
Read how Ensemble Methods for Machine Learning explains local explanation →
Outlier Detection in Python
A local explanation describes why a model produced a particular prediction or score for an individual observation.
- How it is used
- Explains why an individual record received its outlier score, supporting interpretation of otherwise difficult-to-understand detectors.
- How deep it goes
- Describes feature spaces, distributions, and contributing subspaces, while stopping short of recovering the detector’s exact reasoning.
- Examples
- Uses an abalone dataset, two-dimensional KNN models, and a top-scored record repeatedly associated with the Height feature.
- What you can do after
- Diagnose individual outlier scores by identifying feature pairs and spaces most responsible for a record’s anomaly.
Read how Outlier Detection in Python explains local explanation →
Related concepts
- global explanationcontrasts with · A global explanation describes a model’s overall behavior or decision patterns across a range of inputs rather than explaining a single prediction.
- LIMEuses · LIME is a model-agnostic explanation technique that approximates a model near an individual input with an interpretable surrogate to explain its prediction.
- SHAPuses · SHAP is a feature-attribution method that explains model outputs by assigning input features contributions based on Shapley values.
- saliency mapincludes · A saliency map is a visualization indicating how strongly different input features or regions influence a model’s output.
- generationuses · Generation is the computational production of outputs from supplied inputs or context, often by applying a model to produce predictions or new content.
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.