/// <summary>
/// Runs the community experiment.
/// </summary>
/// <param name="trainModel">The train model.</param>
/// <param name="seedInputs">The seed inputs.</param>
/// <param name="communityFeatureSet">The community feature set.</param>
/// <param name="priors">The priors.</param>
/// <param name="thresholdAndNoiseVariance">The threshold and noise variance.</param>
/// <returns>The <see cref="CommunityPosteriors" />.</returns>
public CommunityPosteriors RunCommunityTraining(
CommunityModelBase trainModel,
IList <Inputs> seedInputs,
FeatureSet communityFeatureSet,
CommunityPriors priors,
double thresholdAndNoiseVariance)
{
this.ModelName = trainModel.Name;
// Assume all users have the same feature set type!
this.FeatureSetName = communityFeatureSet.ToString();
trainModel.ConstructModel();
// First train the seeds in offline mode
trainModel.SetObservedValues(
seedInputs.Select(ia => ia.Train.Instances).ToList(),
communityFeatureSet,
InputMode.CommunityTraining,
priors);
var communityResults = new Results();
// Now infer the community posteriors
trainModel.DoInference(communityFeatureSet, seedInputs.Select(ia => ia.UserName), ref communityResults);
this.CommunityResults.Add(communityResults);
return(communityResults.CommunityPosteriors);
}
/// <summary>
/// Runs the personalisation.
/// </summary>
/// <param name="trainModel">The train model.</param>
/// <param name="testModel">The test model.</param>
/// <param name="inputs">The inputs.</param>
/// <param name="priors">The priors.</param>
/// <param name="point">The point.</param>
/// <param name="thresholdAndNoiseVariance">The threshold and noise variance.</param>
/// <param name="limit">The limit.</param>
/// <param name="testMode">The test mode.</param>
private void RunPersonalisation(
CommunityModelBase trainModel,
CommunityModelBase testModel,
Inputs inputs,
CommunityPriors priors,
Func <CommunityPosteriors, CommunityPriors> priorsFromPosteriors,
int limit = -1,
InputMode testMode = InputMode.Validation | InputMode.Testing)
{
trainModel.ConstructModel();
testModel.ConstructModel();
int numTrain = inputs.Train.Count;
if (limit > 0)
{
numTrain = Math.Min(numTrain, limit);
}
if (this.BatchSize == -1)
{
this.BatchSize = numTrain;
}
int numBatches = numTrain / this.BatchSize;
for (int index = 0; index < numBatches; index++)
{
var batchResults = new Results();
var batch = this.GetBatch(inputs, index, InputMode.Training);
if (numBatches > 1)
{
Console.WriteLine(@"{0} batch {1}/{2}, batch size {3}", this.UserName, index + 1, numBatches, batch.Count);
}
trainModel.SetObservedValues(new[] { batch }, inputs.FeatureSet, InputMode.Training, priors);
try
{
trainModel.DoInference(inputs.FeatureSet, new[] { inputs.UserName }, ref batchResults);
}
catch (Exception e)
{
Console.WriteLine(e.Message);
return;
}
// Set the priors to be the posteriors from the previous run
priors = priorsFromPosteriors(batchResults.CommunityPosteriors);
if (this.Mode == ExperimentMode.Online)
{
// clear posteriors to save memory
batchResults.Posteriors = null;
}
// Validation and Test only
testModel.Apply(inputs, inputs.FeatureSet, priors, ref batchResults, testMode);
this.CommunityResults.Add(batchResults);
this.Metrics.Add(new Metrics(inputs, batchResults, ExperimentMode.Online));
this.Count++;
}
}