Bundle
Core machine learning types: the seedable Random generator, Matrix2D linear algebra and activations, matrix reference holders, the NeuralNetwork and CSV-backed MatrixReader. Compile with -lib ml.
ScoreMetric
A statistic over scored instances, so it can be resampled without Bootstrap knowing which one it is. An interface rather than a function reference, matching ScoreModel and Compare elsewhere in the bundle. A metric MUST return a negative value when the sample cannot support it -- AUC over a resample that drew only one class, for example. Bootstrap discards those rather than averaging them in. Returning 0.0 instead would be read as a real score of zero and pull the interval down.
Operations
Compute # virtual
Computes the statistic.
method : public : virtual : Compute(scores:Float[], actuals:Bool[]) ~ FloatParameters
| Name | Type | Description |
|---|---|---|
| scores | Float | one score per instance; higher means more positive |
| actuals | Bool | true where the instance really is positive |
Return
| Type | Description |
|---|---|
| Float | the value, or a negative number when the sample cannot support it |