func (regressor *linearRegressor) Train(trainingData dataset.Dataset) error {
if !trainingData.AllFeaturesFloats() {
return linearerrors.NewNonFloatFeaturesError()
}
if !trainingData.AllTargetsFloats() {
return linearerrors.NewNonFloatTargetsError()
}
if trainingData.NumTargets() != 1 {
return linearerrors.NewInvalidNumberOfTargetsError(trainingData.NumTargets())
}
if trainingData.NumFeatures() == 0 {
return linearerrors.NewNoFeaturesError()
}
estimator, err := gradientdescentestimator.NewGradientDescentParameterEstimator(
defaultLearningRate,
defaultPrecision,
defaultMaxIterations,
gradientdescentestimator.LinearModelLeastSquaresLossGradient,
)
if err != nil {
return linearerrors.NewEstimatorConstructionError(err)
}
err = estimator.Train(trainingData)
if err != nil {
return linearerrors.NewEstimatorTrainingError(err)
}
coefficients, err := estimator.Estimate(defaultInitialCoefficientEstimate(trainingData.NumFeatures()))
if err != nil {
return linearerrors.NewEstimatorEstimationError(err)
}
regressor.coefficients = coefficients
return nil
}
"math"
"math/rand"
. "github.com/onsi/ginkgo"
. "github.com/onsi/gomega"
)
var _ = Describe("Gradient Descent Parameter Estimation", func() {
var lossGradient gradientdescentestimator.ParameterizedLossGradient
Describe("NewGradientDescentParameterEstimator", func() {
Context("Given negative learning rate", func() {
It("Returns an error", func() {
_, err := gradientdescentestimator.NewGradientDescentParameterEstimator(
-0.3,
0.3,
100,
lossGradient,
)
Ω(err).Should(BeAssignableToTypeOf(gdeErrors.InvalidGDPEInitializationValuesError{}))
})
})
Context("Given zero learning rate", func() {
It("Returns an error", func() {
_, err := gradientdescentestimator.NewGradientDescentParameterEstimator(
0,
0.3,
100,
lossGradient,
)
. "github.com/onsi/ginkgo"
. "github.com/onsi/gomega"
)
var _ = Describe("Linear Model Least Squares Parameter Estimation", func() {
var estimator parameterestimator.ParameterEstimator
var trueParameters []float64
Context("Given reasonable learning rate, precision, maximum number of iterations, and training set", func() {
BeforeEach(func() {
trueParameters = []float64{2, -3, 4}
var err error
estimator, err = gradientdescentestimator.NewGradientDescentParameterEstimator(
0.001,
0.000005,
1000,
gradientdescentestimator.LinearModelLeastSquaresLossGradient,
)
Ω(err).ShouldNot(HaveOccurred())
columnTypes, err := columntype.StringsToColumnTypes([]string{"1.0", "1.0", "1.0"})
Ω(err).ShouldNot(HaveOccurred())
trainingSet := dataset.NewDataset([]int{0, 1}, []int{2}, columnTypes)
for i := 0; i < 20; i++ {
for j := 0; j < 20; j++ {
x0 := -1.9 + 0.2*float64(i)
x1 := -1.9 + 0.2*float64(j)
y := trueParameters[0]*x0 + trueParameters[1]*x1 + trueParameters[2] + 0.1*rand.NormFloat64()
err = trainingSet.AddRowFromStrings([]string{