29 lines
855 B
ObjectPascal
29 lines
855 B
ObjectPascal
|
// В этом примере строится confusion matrix для классификации Iris.
|
|||
|
|
|
|||
|
|
uses MLABC, PlotML;
|
|||
|
|
|
|||
|
|
begin
|
|||
|
|
var ds := Datasets.Iris;
|
|||
|
|
|
|||
|
|
var (train, test) := ds.StratifiedTrainTestSplit(0.2, 42);
|
|||
|
|
|
|||
|
|
var Xtrain := train.Data.ToMatrix(train.Features);
|
|||
|
|
var Xtest := test.Data.ToMatrix(test.Features);
|
|||
|
|
|
|||
|
|
var encoder := new LabelEncoder;
|
|||
|
|
var ytrain := encoder.FitTransform(train);
|
|||
|
|
var ytest := encoder.Transform(test);
|
|||
|
|
|
|||
|
|
var model := new LogisticRegression;
|
|||
|
|
model.Fit(Xtrain, ytrain);
|
|||
|
|
|
|||
|
|
var pred := model.Predict(Xtest);
|
|||
|
|
var cm := new ConfusionMatrix(ytest, pred);
|
|||
|
|
var acc := Metrics.Accuracy(ytest, pred);
|
|||
|
|
|
|||
|
|
Plot.ConfusionMatrix(cm, encoder.Classes, normalize := MatrixNormalization.Rows);
|
|||
|
|
Plot.Title := $'Iris: LogisticRegression, accuracy = {acc:F3}';
|
|||
|
|
Plot.XLabel := 'Predicted';
|
|||
|
|
Plot.YLabel := 'Actual';
|
|||
|
|
end.
|