Skip to main contentCrate renegade_ml
Source - vptree
- Vantage-point tree for metric-space nearest neighbor search.
- Dispersion
- Answers “do these neighbors agree?” — a companion to a weighted mean.
- ExtrapolatedPrediction
- Result of extrapolating neighbor outputs to distance=0.
- FeatureDiagnostics
- Diagnostics for a single feature’s learned metric.
- LearnedMetric
- A learned distance metric that transforms features into “effect space” before
computing distances.
- ModelDiagnostics
- Snapshot of the model’s current state for diagnostics/dashboards.
- Neighbor
- A single nearest neighbor result.
- NeighborDetail
- Detail about a single neighbor in a prediction.
- Neighbors
- A set of nearest neighbors, sorted by distance.
- OutputStats
- Statistics about the output values in the training set.
- PredictionDiagnostics
- Diagnostics for a single prediction query.
- Renegade
- The core learner. Stores labeled training data and answers queries via KNN.
- Shrinkage
- Result of
Dispersion::shrink_toward: a local mean pulled toward a
prior by an amount that depends on how much the local evidence is worth
trusting.
- DataPoint
- User implements this trait to define how distances are computed between data points.