7eaddd9a54
ML - устранение неточностей и багов ML - оптимизация производительности DecisionTreeRegressor.Fit, RandomForestRegressor.Fit ML - тесты
20 lines
568 B
ObjectPascal
20 lines
568 B
ObjectPascal
uses MLABC;
|
|
uses TestHelpers in '..\TestHelpers.pas';
|
|
|
|
begin
|
|
var ds := Datasets.Iris;
|
|
var X := ds.Data.ToMatrix(ds.Features);
|
|
|
|
var classes: array of string;
|
|
var y := ds.Data.EncodeLabels(ds.Target, classes);
|
|
|
|
var model := new GradientBoostingClassifier(20, learningRate := 0.1, maxDepth := 3, seed := 1);
|
|
model.Fit(X, y);
|
|
|
|
var proba := model.PredictProba(X);
|
|
|
|
Check(proba.RowCount = X.RowCount, 'Probability row count mismatch');
|
|
Check(proba.ColCount = classes.Length, 'Probability class count mismatch');
|
|
CheckProbabilityRowsSumToOne(proba);
|
|
end.
|