/// <summary>
/// Infers parameters for a given cold item.
/// </summary>
/// <param name="itemFeatures">The item features.</param>
/// <returns>A distribution over the item parameters.</returns>
public ItemParameterDistribution InferItemParameters(SparseFeatureVector itemFeatures)
{
ItemParameterDistribution result;
if (this.itemFeatureParameterPosteriors.FeatureCount == 0)
{
Debug.Assert(
itemFeatures.FeatureCount == 0,
"The number of item features passed must be equal to the number of item features learned.");
result = this.itemParameterDistributionAverage;
}
else
{
GaussianArray traits;
Gaussian bias;
AlgorithmUtils.AddFeatureContribution(
this.itemParameterDistributionAverage,
this.itemFeatureParameterPosteriors,
itemFeatures,
out traits,
out bias);
result = new ItemParameterDistribution(traits, bias);
}
return(result);
}
/// <summary>
/// Sets up the inference algorithm to operate on a single item given the posteriors over the parameters of the item.
/// </summary>
/// <param name="itemParameterPosteriors">The posteriors over the parameters of the item.</param>
private void SetupItemFromPosteriors(ItemParameterDistribution itemParameterPosteriors)
{
Debug.Assert(itemParameterPosteriors != null, "Valid item parameter posteriors must be provided.");
this.inferenceAlgorithm.ItemCount = 1;
this.inferenceAlgorithm.ItemTraitsPrior = new GaussianMatrix(new[] { itemParameterPosteriors.Traits });
this.inferenceAlgorithm.ItemBiasPrior = new GaussianArray(new[] { itemParameterPosteriors.Bias });
}
/// <summary>
/// Initializes a new instance of the <see cref="ColdUserItemParameterAlgorithm"/> class.
/// </summary>
/// <param name="userFeatureParameterPosteriors">The posteriors over the user feature related parameters learned during community training.</param>
/// <param name="itemFeatureParameterPosteriors">The posteriors over the item feature related parameters learned during community training.</param>
/// <param name="userParameterDistributionAverage">The average trait, bias, and threshold posterior over all users in training.</param>
/// <param name="itemParameterDistributionAverage">The average trait and bias posterior over all items in training.</param>
public ColdUserItemParameterAlgorithm(
FeatureParameterDistribution userFeatureParameterPosteriors,
FeatureParameterDistribution itemFeatureParameterPosteriors,
UserParameterDistribution userParameterDistributionAverage,
ItemParameterDistribution itemParameterDistributionAverage)
{
this.userFeatureParameterPosteriors = userFeatureParameterPosteriors;
this.itemFeatureParameterPosteriors = itemFeatureParameterPosteriors;
this.userParameterDistributionAverage = userParameterDistributionAverage;
this.itemParameterDistributionAverage = itemParameterDistributionAverage;
}
/// <summary>
/// Infers the distribution over the rating which a given user will give to an item.
/// </summary>
/// <param name="userParameterPosteriors">The posteriors over user parameters.</param>
/// <param name="itemParameterPosteriors">The posteriors over item parameters.</param>
/// <returns>The distribution over the rating.</returns>
public Discrete InferRatingDistribution(
UserParameterDistribution userParameterPosteriors, ItemParameterDistribution itemParameterPosteriors)
{
Debug.Assert(userParameterPosteriors != null && itemParameterPosteriors != null, "A valid posteriors must be provided.");
Debug.Assert(
userParameterPosteriors.Traits.Count == itemParameterPosteriors.Traits.Count,
"Given posteriors should have the same associated number of traits.");
this.inferenceAlgorithm.TraitCount = userParameterPosteriors.Traits.Count;
this.SetupUserFromPosteriors(userParameterPosteriors);
this.SetupItemFromPosteriors(itemParameterPosteriors);
this.inferenceAlgorithm.ObservationCount = 1;
this.inferenceAlgorithm.UserIds = new[] { 0 };
this.inferenceAlgorithm.ItemIds = new[] { 0 };
this.inferenceAlgorithm.Execute(1);
return(this.inferenceAlgorithm.RatingsMarginal()[0]);
}