public void Linear_Regression_Save_And_Load()
{
var rnd = new Random();
Func<double, double, double> func = (l, r) => l + 2 * r;
LinearRegressionModel model;
var data = new List<ModelItem>();
for (var i = 0; i < 100; i++)
{
var left = rnd.NextDouble(0, 50000);
var right = rnd.NextDouble(0, 50000);
var result = func(left, right);
data.Add(new ModelItem { LeftOperand = left, RightOperand = right, Result = result });
}
var d = Descriptor.Create<ModelItem>();
var g = new LinearRegressionGenerator { Descriptor = d };
var learningModel = Learner.Learn(data, .80, 5, g); // changed from 1000
model = (LinearRegressionModel)learningModel.Model;
Serialize(model);
var loadedModel = Deserialize<LinearRegressionModel>();
Assert.Equal(model.Theta, loadedModel.Theta);
}
public void Linear_Regression_Save_And_Load()
{
var rnd = new Random();
Func <double, double, double> func = (l, r) => l + 2 * r;
LinearRegressionModel model;
var data = new List <ModelItem>();
for (var i = 0; i < 100; i++)
{
var left = rnd.NextDouble(0, 50000);
var right = rnd.NextDouble(0, 50000);
var result = func(left, right);
data.Add(new ModelItem {
LeftOperand = left, RightOperand = right, Result = result
});
}
var d = Descriptor.Create <ModelItem>();
var g = new LinearRegressionGenerator {
Descriptor = d
};
var learningModel = Learner.Learn(data, .80, 5, g); // changed from 1000
model = (LinearRegressionModel)learningModel.Model;
Serialize(model);
var loadedModel = Deserialize <LinearRegressionModel>();
Assert.Equal(model.Theta, loadedModel.Theta);
}
/// <summary>
/// Loads the prediction.
/// </summary>
private static void LoadPrediction()
{
var cupcakeData = Encoding.UTF8.GetString(Resources.CupcakeData);
var cupcakePan = JsonConvert.DeserializeObject <CupcakePan>(cupcakeData);
// Initalize the X model
var generatorX = new LinearRegressionGenerator
{
Descriptor =
Descriptor.For <CupcakeIngredientX>()
.With(x => x.Delay)
.With(x => x.MissileSpeed)
.WithEnumerable(x => x.SourcePosition, 3)
.WithEnumerable(x => x.TargetPosition, 3)
//.WithEnumerable(x => x.Waypoints, 3)
.With(x => x.TargetMoveSpeed)
.With(x => x.Width)
.Learn(x => x.BakedX)
};
var learnedX = Learner.Learn(cupcakePan.X, 0.80, 1000, generatorX);
xmodel = learnedX.Model;
// Initalize the Y model
var generatorY = new LinearRegressionGenerator()
{
Descriptor =
Descriptor.For <CupcakeIngredientY>()
.With(x => x.Delay)
.With(x => x.MissileSpeed)
.WithEnumerable(x => x.SourcePosition, 3)
.WithEnumerable(x => x.TargetPosition, 3)
// .WithEnumerable(x => x.Waypoints, 3)
.With(x => x.TargetMoveSpeed)
.With(x => x.Width)
.Learn(x => x.BakedY)
};
var learnedY = Learner.Learn(cupcakePan.Y, 0.80, 1000, generatorY);
ymodel = learnedY.Model;
Initialized = true;
}
public void Linear_Regression_Test_House_Predictions_Normal()
{
// test house prices based on ft-sq and no# bedrooms
Matrix x = new [, ]
{
{ 2104, 3 },
{ 1600, 3 },
{ 2400, 3 },
{ 1416, 2 },
{ 3000, 4 },
{ 1985, 4 },
{ 1534, 3 },
{ 1427, 3 },
{ 1380, 3 },
{ 1494, 3 }
};
Vector y = new[]
{
399900,
329900,
369000,
232000,
539900,
299900,
314900,
198999,
212000,
242500
};
LinearRegressionGenerator generator = new LinearRegressionGenerator()
{
LearningRate = 0.01, MaxIterations = 400, Lambda = 0, NormalizeFeatures = true
};
var model = generator.Generate(x.Copy(), y.Copy());
var priceEqns = model.Predict(new Vector(new double[] { 1650, 3 }));
// CK 150929: increased due to improvements in optimisation
double actualEqns = 295107.0d;
Almost.Equal(actualEqns, System.Math.Round(priceEqns, 0), 5000);
}
//[Test]
public void Save_And_Load_LinearRegression()
{
Matrix x = new [, ]
{
{ 2104, 3 },
{ 1600, 3 },
{ 2400, 3 },
{ 1416, 2 },
{ 3000, 4 },
{ 1985, 4 },
{ 1534, 3 },
{ 1427, 3 },
{ 1380, 3 },
{ 1494, 3 }
};
Vector y = new[]
{
399900,
329900,
369000,
232000,
539900,
299900,
314900,
198999,
212000,
242500
};
LinearRegressionGenerator generator = new LinearRegressionGenerator()
{
LearningRate = 0.01, MaxIterations = 400, Lambda = 1
};
var model = generator.Generate(x.Copy(), y.Copy()) as LinearRegressionModel;
Serialize(model);
var lmodel = Deserialize <LinearRegressionModel>();
Assert.AreEqual(model.Theta, lmodel.Theta);
}
public void Linear_Regression_Test_House_Predictions_Regularized()
{
// test house prices based on ft-sq and no# bedrooms
Matrix x = new [, ]
{
{ 2104, 3 },
{ 1600, 3 },
{ 2400, 3 },
{ 1416, 2 },
{ 3000, 4 },
{ 1985, 4 },
{ 1534, 3 },
{ 1427, 3 },
{ 1380, 3 },
{ 1494, 3 }
};
Vector y = new[]
{
399900,
329900,
369000,
232000,
539900,
299900,
314900,
198999,
212000,
242500
};
LinearRegressionGenerator generator = new LinearRegressionGenerator()
{
LearningRate = 0.01, MaxIterations = 400, Lambda = 1
};
var model = generator.Generate(x.Copy(), y.Copy());
var priceGrad = model.Predict(new Vector(new double[] { 1650, 3 }));
double actualGrad = 280942d;
Assert.AreEqual(actualGrad, System.Math.Round(priceGrad, 0));
}
public void Linear_Regression_Learner_Test()
{
var datum = new[]
{
new { X = 1.0, Y = 1.0 },
new { X = 2.0, Y = 2.0 },
new { X = 3.0, Y = 1.3 },
new { X = 4.0, Y = 3.75 },
new { X = 5.0, Y = 5.25 }
};
var d = Descriptor.New("DATUM")
.With("X").As(typeof(double))
.Learn("Y").As(typeof(double));
var generator = new LinearRegressionGenerator()
{
Descriptor = d
};
var model = Learner.Learn(datum, 0.9, 5, generator);
Assert.True(0.8 >= model.Accuracy);
}
/// <summary>
/// Loads the prediction.
/// </summary>
private static void LoadPrediction()
{
var cupcakeData = Encoding.UTF8.GetString(Resources.CupcakeData);
var cupcakePan = JsonConvert.DeserializeObject<CupcakePan>(cupcakeData);
// Initalize the X model
var generatorX = new LinearRegressionGenerator
{
Descriptor =
Descriptor.For<CupcakeIngredientX>()
.With(x => x.Delay)
.With(x => x.MissileSpeed)
.WithEnumerable(x => x.SourcePosition, 3)
.WithEnumerable(x => x.TargetPosition, 3)
//.WithEnumerable(x => x.Waypoints, 3)
.With(x => x.TargetMoveSpeed)
.With(x => x.Width)
.Learn(x => x.BakedX)
};
var learnedX = Learner.Learn(cupcakePan.X, 0.80, 1000, generatorX);
xmodel = learnedX.Model;
// Initalize the Y model
var generatorY = new LinearRegressionGenerator()
{
Descriptor =
Descriptor.For<CupcakeIngredientY>()
.With(x => x.Delay)
.With(x => x.MissileSpeed)
.WithEnumerable(x => x.SourcePosition, 3)
.WithEnumerable(x => x.TargetPosition, 3)
// .WithEnumerable(x => x.Waypoints, 3)
.With(x => x.TargetMoveSpeed)
.With(x => x.Width)
.Learn(x => x.BakedY)
};
var learnedY = Learner.Learn(cupcakePan.Y, 0.80, 1000, generatorY);
ymodel = learnedY.Model;
Initialized = true;
}
