private IDistribution createGaussianMixture(double mean, double std, double mean2, double std2)
{
IDistribution d1 = create1DGaussian(mean, std);
IDistribution d2 = create1DGaussian(mean2, std2);
int dim = 1;
string [] names = new string [1] {
"x"
};
double [] mins = new double [1] {
0.00
};
double [] maxs = new double [1] {
100.0
};
IBlauSpace s = BlauSpace.create(dim, names, mins, maxs);
Mixture d = new Mixture(s);
d.Add(d1, 0.75);
d.Add(d2, 0.25);
d.DistributionComplete();
return(d);
}
public void DistributionSpaceIterator_GaussianMixtureTest()
{
Console.WriteLine("DistributionSpaceIterator_GaussianMixtureTest");
double mean = 70.0;
double std = 1.0;
IDistribution d1 = create1DGaussian(mean, std);
double mean2 = 20.0;
double std2 = 1.0;
IDistribution d2 = create1DGaussian(mean2, std2);
int dim = 1;
string [] names = new string [1] {
"x"
};
double [] mins = new double [1] {
0.00
};
double [] maxs = new double [1] {
100.0
};
IBlauSpace s = BlauSpace.create(dim, names, mins, maxs);
Mixture d = new Mixture(s);
d.Add(d1, 0.75);
d.Add(d2, 0.25);
d.DistributionComplete();
SingletonLogger.Instance().DebugLog(typeof(dist_tests), "original distribution: " + d);
DistributionSpace ds = new DistributionSpace(d);
int [] steps = new int[ds.ParamSpace.Dimension];
for (int N = 3; N <= 5; N++)
{
for (int i = 0; i < ds.ParamSpace.Dimension; i++)
{
steps[i] = N;
}
IDistributionSpaceIterator it = ds.iterator(steps);
int count = 0;
int validCt = 0;
foreach (IDistribution diter in it)
{
if (diter.IsValid())
{
validCt++;
SingletonLogger.Instance().DebugLog(typeof(dist_tests), "iterator distribution: " + diter);
}
count++;
}
Assert.AreEqual((N + 1) * (N + 1) * (N + 1) * (N + 1) * (N + 1) * (N + 1), count);
SingletonLogger.Instance().InfoLog(typeof(dist_tests), "N=" + N + " valid distributions: " + validCt + " / total: " + count);
}
}
public Mixture<VectorGaussian> ToMixture()
{
Mixture<VectorGaussian> mixture = new Mixture<VectorGaussian>();
for (int i = 0; i < weights.Count; ++i)
mixture.Add(
VectorGaussian.FromMeanAndVariance(
MicrosoftResearch.Infer.Maths.Vector.FromArray(this.means[i]),
new PositiveDefiniteMatrix(JaggedArrayToMatrix(this.variances[i]))),
this.weights[i]);
return mixture;
}
public Mixture <VectorGaussian> ToMixture()
{
Mixture <VectorGaussian> mixture = new Mixture <VectorGaussian>();
for (int i = 0; i < weights.Count; ++i)
{
mixture.Add(
VectorGaussian.FromMeanAndVariance(
MicrosoftResearch.Infer.Maths.Vector.FromArray(this.means[i]),
new PositiveDefiniteMatrix(JaggedArrayToMatrix(this.variances[i]))),
this.weights[i]);
}
return(mixture);
}
public static Mixture <VectorGaussian> Fit(MicrosoftResearch.Infer.Maths.Vector[] data, int componentCount, int retryCount, double tolerance = 1e-4)
{
Debug.Assert(data != null);
Debug.Assert(data.Length > componentCount * 3);
Debug.Assert(componentCount > 1);
Debug.Assert(retryCount >= 0);
int dimensions = data[0].Count;
// Find point boundary
MicrosoftResearch.Infer.Maths.Vector min = data[0].Clone();
MicrosoftResearch.Infer.Maths.Vector max = min.Clone();
for (int i = 1; i < data.Length; ++i)
{
Debug.Assert(dimensions == data[i].Count);
for (int j = 0; j < dimensions; ++j)
{
min[j] = Math.Min(min[j], data[i][j]);
max[j] = Math.Max(max[j], data[i][j]);
}
}
// Initialize solution
MicrosoftResearch.Infer.Maths.Vector[] means = new MicrosoftResearch.Infer.Maths.Vector[componentCount];
PositiveDefiniteMatrix[] covariances = new PositiveDefiniteMatrix[componentCount];
for (int i = 0; i < componentCount; ++i)
{
GenerateRandomMixtureComponent(min, max, out means[i], out covariances[i]);
}
double[] weights = Enumerable.Repeat(1.0 / componentCount, componentCount).ToArray();
// EM algorithm for GMM
double[,] expectations = new double[data.Length, componentCount];
double lastEstimate;
const double negativeInfinity = -1e+20;
bool convergenceDetected;
double currentEstimate = negativeInfinity;
do
{
lastEstimate = currentEstimate;
convergenceDetected = false;
// E-step: estimate expectations on hidden variables
for (int i = 0; i < data.Length; ++i)
{
double sum = 0;
for (int j = 0; j < componentCount; ++j)
{
expectations[i, j] =
Math.Exp(VectorGaussian.GetLogProb(data[i], means[j], covariances[j])) * weights[j];
sum += expectations[i, j];
}
for (int j = 0; j < componentCount; ++j)
{
expectations[i, j] /= sum;
}
}
// M-step:
// Re-estimate means
for (int j = 0; j < componentCount; ++j)
{
means[j] = MicrosoftResearch.Infer.Maths.Vector.Zero(dimensions);
double sum = 0;
for (int i = 0; i < data.Length; ++i)
{
means[j] += data[i] * expectations[i, j];
sum += expectations[i, j];
}
means[j] *= 1.0 / sum;
}
// Re-estimate covariances
for (int j = 0; j < componentCount; ++j)
{
Matrix covariance = new Matrix(dimensions, dimensions);
double sum = 0;
for (int i = 0; i < data.Length; ++i)
{
MicrosoftResearch.Infer.Maths.Vector dataDiff = data[i] - means[j];
covariance += dataDiff.Outer(dataDiff) * expectations[i, j];
sum += expectations[i, j];
}
covariance *= 1.0 / sum;
covariances[j] = new PositiveDefiniteMatrix(covariance);
if (covariances[j].LogDeterminant() < -30)
{
DebugConfiguration.WriteDebugText("Convergence detected for component {0}", j);
if (retryCount == 0)
{
throw new InvalidOperationException("Can't fit GMM. Retry number exceeded.");
}
retryCount -= 1;
GenerateRandomMixtureComponent(min, max, out means[j], out covariances[j]);
DebugConfiguration.WriteDebugText("Component {0} regenerated", j);
convergenceDetected = true;
}
}
if (convergenceDetected)
{
currentEstimate = negativeInfinity;
continue;
}
// Re-estimate weights
double expectationSum = 0;
for (int j = 0; j < componentCount; ++j)
{
weights[j] = 0;
for (int i = 0; i < data.Length; ++i)
{
weights[j] += expectations[i, j];
expectationSum += expectations[i, j];
}
}
for (int j = 0; j < componentCount; ++j)
{
weights[j] /= expectationSum;
}
// Compute likelihood estimate
currentEstimate = 0;
for (int i = 0; i < data.Length; ++i)
{
for (int j = 0; j < componentCount; ++j)
{
currentEstimate +=
expectations[i, j] * (VectorGaussian.GetLogProb(data[i], means[j], covariances[j]) + Math.Log(weights[j]));
}
}
DebugConfiguration.WriteDebugText("L={0:0.000000}", currentEstimate);
} while (convergenceDetected || (currentEstimate - lastEstimate > tolerance));
Mixture <VectorGaussian> result = new Mixture <VectorGaussian>();
for (int j = 0; j < componentCount; ++j)
{
result.Add(VectorGaussian.FromMeanAndVariance(means[j], covariances[j]), weights[j]);
}
DebugConfiguration.WriteDebugText("GMM successfully fitted.");
return(result);
}
public static void Main1DMixture(string[] args)
{
Console.WriteLine("console_tests");
LoggerInitialization.SetThreshold(typeof(console_tests.MainClass), LogLevel.Debug);
double mean = 70.0;
double std = 1.0;
IDistribution d1 = create1DGaussian(mean, std);
double mean2 = 20.0;
double std2 = 1.0;
IDistribution d2 = create1DGaussian(mean2, std2);
int dim = 1;
string [] names = new string [1] {
"x"
};
double [] mins = new double [1] {
0.00
};
double [] maxs = new double [1] {
100.0
};
IBlauSpace s = BlauSpace.create(dim, names, mins, maxs);
Mixture d = new Mixture(s);
d.Add(d1, 0.75);
d.Add(d2, 0.25);
d.DistributionComplete();
SingletonLogger.Instance().DebugLog(typeof(console_tests.MainClass), "original distribution: " + d);
Console.WriteLine("original distribution: " + d);
DistributionSpace ds = new DistributionSpace(d);
int [] steps = new int[ds.ParamSpace.Dimension];
// N = subdivisions of each of the parameter values
for (int N = 3; N <= 6; N++)
{
for (int i = 0; i < ds.ParamSpace.Dimension; i++)
{
steps[i] = N;
}
IDistributionSpaceIterator it = ds.iterator(steps);
int count = 0;
int validCt = 0;
foreach (IDistribution diter in it)
{
if (diter.IsValid())
{
validCt++;
SingletonLogger.Instance().DebugLog(typeof(console_tests.MainClass), "iterator distribution: " + diter);
Console.WriteLine("valid distribution: " + diter);
}
else
{
Console.WriteLine("invalid distribution: " + diter);
}
count++;
}
Console.WriteLine("# of valid distributions: " + validCt);
Console.WriteLine("# of total distributions: " + count);
Assert.AreEqual((N + 1) * (N + 1) * (N + 1) * (N + 1) * (N + 1) * (N + 1), count);
SingletonLogger.Instance().InfoLog(typeof(console_tests.MainClass), "N=" + N + " valid distributions: " + validCt + " / total: " + count);
}
}
public static void MakeMixture_Main(string[] args)
{
Console.WriteLine("MakeMixture");
// Command line parsing
Arguments CommandLine = new Arguments(args);
bool err = false;
string errString = "";
string file1 = "unassigned";
string file2 = "unassigned";
string outfile = "unassigned";
// Look for specific arguments values and display
// them if they exist (return null if they don't)
if (CommandLine["file1"] != null)
{
file1 = CommandLine["file1"];
if (!File.Exists(file1))
{
errString += ("The specified 'file1' was not found: " + file1 + " ");
err = true;
}
}
else
{
errString += ("The 'file1' was not specified. ");
err = true;
}
double weight1 = -1.0;
// Look for specific arguments values and display
// them if they exist (return null if they don't)
if (CommandLine["weight1"] != null)
{
try {
weight1 = Double.Parse(CommandLine["weight1"]);
if ((weight1 < 0.0) || (weight1 > 1.0))
{
errString += ("The specified 'weight1' was not in the range [0,1]. ");
err = true;
}
}
catch (Exception) {
errString += ("The specified 'weight1' was not valid. ");
err = true;
}
}
else
{
errString += ("The 'weight1' was not specified. ");
err = true;
}
if (CommandLine["file2"] != null)
{
file2 = CommandLine["file2"];
if (!File.Exists(file2))
{
errString += ("The specified 'file2' was not found: " + file2 + " ");
err = true;
}
}
else
{
errString += ("The 'file2' was not specified. ");
err = true;
}
if (CommandLine["outfile"] != null)
{
outfile = CommandLine["outfile"];
}
else
{
errString += ("The 'outfile' was not specified. ");
err = true;
}
if (err)
{
Console.Out.WriteLine("Arguments parsing failed.");
Console.Out.WriteLine(" " + errString);
}
else
{
Console.Out.WriteLine("Arguments parsing successful.");
Console.Out.WriteLine(" file1 = " + file1);
Console.Out.WriteLine(" weight1 = " + weight1);
Console.Out.WriteLine(" file2 = " + file2);
Console.Out.WriteLine(" weight2 = " + (1 - weight1));
Console.Out.WriteLine(" outfile = " + outfile);
SoapFormatter formatter = new SoapFormatter();
FileStream fs = new FileStream(file1, FileMode.Open);
IDistribution d1 = (IDistribution)formatter.Deserialize(fs);
fs.Close();
fs = new FileStream(file2, FileMode.Open);
IDistribution d2 = (IDistribution)formatter.Deserialize(fs);
fs.Close();
if (!BlauSpace.contains(d1.SampleSpace, d2.SampleSpace) || !BlauSpace.contains(d2.SampleSpace, d1.SampleSpace))
{
Console.Out.WriteLine("The sample spaces of the two distributions are not identical intersection.");
Console.Out.WriteLine(" d1: " + d1);
Console.Out.WriteLine(" d2: " + d2);
Console.Out.WriteLine("The mixture must be constructed over identical sample spaces.");
Console.Out.WriteLine("This is a fatal error, preventing the construction of the mixture distribution.");
}
else
{
int dim3 = d1.SampleSpace.Dimension;
string [] names3 = new string [dim3];
double [] mins3 = new double [dim3];
double [] maxs3 = new double [dim3];
for (int i = 0; i < d1.SampleSpace.Dimension; i++)
{
names3[i] = d1.SampleSpace.getAxis(i).Name;
mins3[i] = d1.SampleSpace.getAxis(i).MinimumValue;
maxs3[i] = d1.SampleSpace.getAxis(i).MaximumValue;
}
IBlauSpace s3 = BlauSpace.create(dim3, names3, mins3, maxs3);
Mixture d3 = new Mixture(s3);
d3.Add(d1, weight1);
d3.Add(d2, 1.0 - weight1);
d3.DistributionComplete();
Console.Out.WriteLine("Distribution: " + d3);
fs = new FileStream(outfile, FileMode.Create);
formatter.Serialize(fs, d3);
fs.Close();
}
}
}
