private Variable <Vector> InitialiseWeights(
DistributionType distributionType,
DistributionName distributionName,
int dimension,
string[] hyperParameters)
{
switch (distributionName)
{
case DistributionName.GaussianDefault:
return(Variable.Random(new VectorGaussian(
Vector.Zero(dimension),
PositiveDefiniteMatrix.Identity(dimension))).Named(
"w." + distributionType.ToString()));
case DistributionName.GaussianInit:
return(Variable <Vector> .Random(
Variable.New <VectorGaussian>().Named(
"w." + distributionType.ToString())));
default:
TraceListeners.Log(TraceEventType.Error, 0,
"Invalid distribution name: " + distributionName.ToString(), true, true);
return(null);
}
}
public override void Train(
string outputModelFileName,
int iterationCount,
bool computeModelEvidence,
int batchCount,
DistributionName distributionName,
InferenceAlgorithm inferenceEngineAlgorithm,
double noise)
{
TraceListeners.Log(TraceEventType.Warning, 0,
"Advanced setting will not be used: " +
"distributionName, inferenceEngineAlgorithm & noise.", false, true);
// Validate
_validate.Train(
outputModelFileName: outputModelFileName,
iterationCount: iterationCount,
batchCount: batchCount);
Train(
outputModelFileName,
iterationCount,
computeModelEvidence,
batchCount);
}
public override void Train(
string outputModelFileName,
int iterationCount,
bool computeModelEvidence,
int batchCount,
DistributionName distributionName,
InferenceAlgorithm inferenceEngineAlgorithm,
double noise)
{
TraceListeners.Log(TraceEventType.Warning, 0,
"Advanced setting will not be used: " +
"batchCount.", false, true);
// Validate
_validate.Train(
outputModelFileName: outputModelFileName,
iterationCount: iterationCount,
batchCount: batchCount);
// The inference engine
_engine[DistributionType.Prior] = SetInferenceEngine(
inferenceEngineAlgorithm, iterationCount);
// Initialise prior weights
_w[DistributionType.Prior] = InitialiseWeights(
distributionType: DistributionType.Prior,
distributionName: distributionName,
dimension: _numFeatures,
hyperParameters: null);
// BPM
BayesPointMachine(_x, _y, _w[DistributionType.Prior], noise);
}
public virtual void Predict(
string inputModelFileName,
DistributionName distributionName = DistributionName.Null,
InferenceAlgorithm inferenceEngineAlgorithm = InferenceAlgorithm.Null,
double noise = 0)
{
TraceListeners.Log(TraceEventType.Information, 0,
"Predict...", false, true);
_model.Predict(
inputModelFileName,
distributionName,
inferenceEngineAlgorithm,
noise);
}
public virtual IDictionary <string, double> PredictInstance(
string inputModelFileName,
DistributionName distributionName,
InferenceAlgorithm inferenceEngineAlgorithm,
int instance,
double noise = 0)
{
TraceListeners.Log(TraceEventType.Information, 0,
"PredictInstance...", false, true);
return(_model.PredictInstance(
inputModelFileName,
distributionName,
inferenceEngineAlgorithm,
instance,
noise));
}
private Variable <Vector> InferWeights(
DistributionType distributionType,
DistributionName distributionName,
string[] hyperParameters)
{
switch (distributionName)
{
case DistributionName.GaussianDefault:
VectorGaussian wObserved = _engine[distributionType].Infer <VectorGaussian>(_w[DistributionType.Prior]);
return(Variable.Random(wObserved).Named("w." + distributionType.ToString()));
default:
TraceListeners.Log(TraceEventType.Error, 0,
"Invalid distribution name: " + distributionName.ToString(), true, true);
return(null);
}
}
public override void Predict(
string inputModelFileName,
DistributionName distributionName,
InferenceAlgorithm inferenceEngineAlgorithm,
double noise)
{
// Validate
// _validate.Predict(inputModelFileName);
// Initialise
List <IDictionary <string, double> > yPredicDistrib = new List <IDictionary <string, double> >();
List <string> yPredicLabel = new List <string>();
VariableArray <bool> yTest = Variable.Array <bool>(new Range(_numObservations)).
Named("y." + _availableDatasetName.ToString()); // to do: correct the size
Bernoulli[] yPredic;
// The inference engine
_engine[DistributionType.Posterior] = SetInferenceEngine(
inferenceEngineAlgorithm, 1);
// Infer postrior weights from training
_w[DistributionType.Posterior] = InferWeights(
distributionType: DistributionType.Posterior,
distributionName: distributionName,
hyperParameters: null);
// BPM
BayesPointMachine(_x, yTest, _w[DistributionType.Posterior], noise);
// predict
// _yPredicLabel = _engine[DistributionType.Posterior].Infer(_y);
yPredic = _engine[DistributionType.Posterior].Infer <Bernoulli[]>(yTest);
for (int i = 0; i < yPredic.Length; i++)
{
yPredicDistrib.Add(new Dictionary <string, double>()
{
{ Convert.ToInt32(yPredic[i].GetProbFalse() > _cutoffPoint).ToString(), yPredic[i].GetProbFalse() },
{ Convert.ToInt32(yPredic[i].GetProbTrue() > _cutoffPoint).ToString(), yPredic[i].GetProbTrue() }
});
yPredicLabel.Add(Convert.ToInt32(yPredic[i].GetProbTrue() > _cutoffPoint).ToString());
}
_yPredicDistrib = yPredicDistrib;
_yPredicLabel = yPredicLabel;
}
public virtual void Train(
string outputModelFileName,
int iterationCount,
bool computeModelEvidence,
int batchCount,
DistributionName distributionName = DistributionName.Null,
InferenceAlgorithm inferenceEngineAlgorithm = InferenceAlgorithm.Null,
double noise = 0)
{
TraceListeners.Log(TraceEventType.Information, 0,
"Train...", false, true);
_model.Train(
outputModelFileName,
iterationCount,
computeModelEvidence,
batchCount,
distributionName,
inferenceEngineAlgorithm,
noise);
}
public override IDictionary <string, double> PredictInstance(
string inputModelFileName,
DistributionName distributionName,
InferenceAlgorithm inferenceEngineAlgorithm,
int instance,
double noise)
{
TraceListeners.Log(TraceEventType.Warning, 0,
"Advanced setting will not be used: " +
"distributionName, inferenceEngineAlgorithm & noise.", false, true);
// Validate
_validate.PredictInstance(
inputModelFileName: inputModelFileName,
instance: instance,
numObservations: _numObservations);
IBayesPointMachineClassifier <
IList <Vector>, int, IList <string>, string, IDictionary <string, double>,
BayesPointMachineClassifierTrainingSettings,
BinaryBayesPointMachineClassifierPredictionSettings <string> > classifier = null;
// Load model
if (string.IsNullOrEmpty(inputModelFileName))
{
classifier =
BayesPointMachineClassifier.LoadBinaryClassifier <
IList <Vector>, int, IList <string>, string, IDictionary <string, double> >
(inputModelFileName);
}
else
{
classifier = _classifier;
}
IDictionary <string, double> yPredicted =
classifier.PredictDistribution(instance, _x);
// string yPredicLabel = classifier.Predict(instance, _x);
return(yPredicted);
}
public override IDictionary <string, double> PredictInstance(
string inputModelFileName,
DistributionName distributionName,
InferenceAlgorithm inferenceEngineAlgorithm,
int instance,
double noise)
{
// Validate
// _validate.Predict(inputModelFileName);
// Initialise
Dictionary <string, double> yPredicDistrib = new Dictionary <string, double>();
VariableArray <bool> yTest = Variable.Array <bool>(new Range(1)).
Named("y." + _availableDatasetName.ToString()); // to do: correct the size
Bernoulli yPredic;
// The inference engine
_engine[DistributionType.Posterior] = SetInferenceEngine(
inferenceEngineAlgorithm, 1);
// Infer postrior weights from training
_w[DistributionType.Posterior] = InferWeights(
distributionType: DistributionType.Posterior,
distributionName: distributionName,
hyperParameters: null);
// BPM
BayesPointMachine(_x[instance], yTest[0], _w[DistributionType.Posterior], noise);
// predict
// _yPredicLabel = _engine[DistributionType.Posterior].Infer(_y);
yPredic = _engine[DistributionType.Posterior].Infer <Bernoulli>(yTest[0]);
yPredicDistrib = new Dictionary <string, double>()
{
{ Convert.ToInt32(yPredic.GetProbFalse() > _cutoffPoint).ToString(), yPredic.GetProbFalse() },
{ Convert.ToInt32(yPredic.GetProbTrue() > _cutoffPoint).ToString(), yPredic.GetProbTrue() }
};
return(yPredicDistrib);
}
public override void TrainIncremental(
string inputModelFileName,
string outputModelFileName,
int iterationCount,
bool computeModelEvidence,
int batchCount,
DistributionName distributionName,
InferenceAlgorithm inferenceEngineAlgorithm,
double noise)
{
TraceListeners.Log(TraceEventType.Warning, 0,
"Advanced setting will not be used: " +
"distributionName, inferenceEngineAlgorithm & noise.", false, true);
// Validate
_validate.TrainIncremental(
inputModelFileName: inputModelFileName,
outputModelFileName: outputModelFileName,
iterationCount: iterationCount,
batchCount: batchCount);
// Load model
IBayesPointMachineClassifier <
IList <Vector>, int, IList <string>, string, IDictionary <string, double>,
BayesPointMachineClassifierTrainingSettings,
BinaryBayesPointMachineClassifierPredictionSettings <string> > classifier =
BayesPointMachineClassifier.LoadBinaryClassifier <
IList <Vector>, int, IList <string>, string, IDictionary <string, double> >
(inputModelFileName);
// Set settings
classifier.Settings.Training.ComputeModelEvidence = computeModelEvidence;
classifier.Settings.Training.IterationCount = iterationCount;
classifier.Settings.Training.BatchCount = batchCount;
// train
classifier.TrainIncremental(_x, _y);
classifier.Save(outputModelFileName);
}
public override void Predict(
string inputModelFileName,
DistributionName distributionName,
InferenceAlgorithm inferenceEngineAlgorithm,
double noise)
{
TraceListeners.Log(TraceEventType.Warning, 0,
"Advanced setting will not be used: " +
"distributionName, inferenceEngineAlgorithm & noise.", false, true);
// Validate
_validate.Predict(inputModelFileName);
// Define the classifier
IBayesPointMachineClassifier <
IList <Vector>, int, IList <string>, string, IDictionary <string, double>,
BayesPointMachineClassifierTrainingSettings,
BinaryBayesPointMachineClassifierPredictionSettings <string> > classifier = null;
// Load model
if (string.IsNullOrEmpty(inputModelFileName))
{
classifier =
BayesPointMachineClassifier.LoadBinaryClassifier <
IList <Vector>, int, IList <string>, string, IDictionary <string, double> >
(inputModelFileName);
}
else
{
classifier = _classifier;
}
_validate.ValidatePredict(_x, _x);
_yPredicDistrib = classifier.PredictDistribution(_x);
_yPredicLabel = classifier.Predict(_x);
}
