public void BinaryRecommendationCustomSerializationRegressionTest()
{
// TODO: We need a static class which implements mapping chainings
Rand.Restart(12347);
var positiveOnlyMapping = new PositiveOnlyDataMapping();
var negativeDataGeneratorMapping = positiveOnlyMapping.WithGeneratedNegativeData(1.0);
using (Stream stream = new MemoryStream())
{
negativeDataGeneratorMapping.SaveForwardCompatibleAsBinary(stream);
stream.Seek(0, SeekOrigin.Begin);
using (var reader = new WrappedBinaryReader(new BinaryReader(stream)))
{
var deserializedMapping =
new NegativeDataGeneratorMapping <PositiveOnlyDataset, Tuple <string, string>, string, string, FeatureProvider, Vector>(reader, positiveOnlyMapping);
var recommender = MatchboxRecommender.Create(deserializedMapping);
recommender.Train(this.positiveOnlyDataset, this.featureProvider); // must not throw
foreach (var instance in positiveOnlyMapping.GetInstances(this.positiveOnlyDataset))
{
var user = positiveOnlyMapping.GetUser(this.positiveOnlyDataset, instance);
var item = positiveOnlyMapping.GetItem(this.positiveOnlyDataset, instance);
Assert.Equal(1, recommender.Predict(user, item, this.featureProvider));
}
}
}
}
/// <summary>
/// Runs the module.
/// </summary>
/// <param name="args">The command line arguments for the module.</param>
/// <param name="usagePrefix">The prefix to print before the usage string.</param>
/// <returns>True if the run was successful, false otherwise.</returns>
public override bool Run(string[] args, string usagePrefix)
{
string datasetFile = string.Empty;
string trainedModelFile = string.Empty;
string predictionsFile = string.Empty;
var parser = new CommandLineParser();
parser.RegisterParameterHandler("--data", "FILE", "Dataset to make predictions for", v => datasetFile = v, CommandLineParameterType.Required);
parser.RegisterParameterHandler("--model", "FILE", "File with trained model", v => trainedModelFile = v, CommandLineParameterType.Required);
parser.RegisterParameterHandler("--predictions", "FILE", "File with generated predictions", v => predictionsFile = v, CommandLineParameterType.Required);
if (!parser.TryParse(args, usagePrefix))
{
return(false);
}
RecommenderDataset testDataset = RecommenderDataset.Load(datasetFile);
var trainedModel = MatchboxRecommender.Load <RecommenderDataset, User, Item, RatingDistribution, DummyFeatureSource>(trainedModelFile);
IDictionary <User, IDictionary <Item, int> > predictions = trainedModel.Predict(testDataset);
RecommenderPersistenceUtils.SavePredictedRatings(predictionsFile, predictions);
return(true);
}
/// <summary>
/// Runs the module.
/// </summary>
/// <param name="args">The command line arguments for the module.</param>
/// <param name="usagePrefix">The prefix to print before the usage string.</param>
/// <returns>True if the run was successful, false otherwise.</returns>
public override bool Run(string[] args, string usagePrefix)
{
string datasetFile = string.Empty;
string trainedModelFile = string.Empty;
string predictionsFile = string.Empty;
int maxRelatedItemCount = 5;
int minCommonRatingCount = 5;
int minRelatedItemPoolSize = 5;
var parser = new CommandLineParser();
parser.RegisterParameterHandler("--data", "FILE", "Dataset to make predictions for", v => datasetFile = v, CommandLineParameterType.Required);
parser.RegisterParameterHandler("--model", "FILE", "File with trained model", v => trainedModelFile = v, CommandLineParameterType.Required);
parser.RegisterParameterHandler("--predictions", "FILE", "File with generated predictions", v => predictionsFile = v, CommandLineParameterType.Required);
parser.RegisterParameterHandler("--max-items", "NUM", "Maximum number of related items for a single item; defaults to 5", v => maxRelatedItemCount = v, CommandLineParameterType.Optional);
parser.RegisterParameterHandler("--min-common-users", "NUM", "Minimum number of users that the query item and the related item should have been rated by in common; defaults to 5", v => minCommonRatingCount = v, CommandLineParameterType.Optional);
parser.RegisterParameterHandler("--min-pool-size", "NUM", "Minimum size of the related item pool for a single item; defaults to 5", v => minRelatedItemPoolSize = v, CommandLineParameterType.Optional);
if (!parser.TryParse(args, usagePrefix))
{
return(false);
}
RecommenderDataset testDataset = RecommenderDataset.Load(datasetFile);
var trainedModel = MatchboxRecommender.Load <RecommenderDataset, User, Item, RatingDistribution, DummyFeatureSource>(trainedModelFile);
var evaluator = new RecommenderEvaluator <RecommenderDataset, User, Item, int, int, RatingDistribution>(
Mappings.StarRatingRecommender.ForEvaluation());
IDictionary <Item, IEnumerable <Item> > relatedItems = evaluator.FindRelatedItemsRatedBySameUsers(
trainedModel, testDataset, maxRelatedItemCount, minCommonRatingCount, minRelatedItemPoolSize);
RecommenderPersistenceUtils.SaveRelatedItems(predictionsFile, relatedItems);
return(true);
}
/// <summary>
/// Creates an instance of the Matchbox-based recommender using the settings from this configuration element.
/// </summary>
/// <param name="mapping">The mapping for the recommender being created.</param>
/// <returns>An instance of the Matchbox-based recommender.</returns>
public override IRecommender <SplitInstanceSource <RecommenderDataset>, User, Item, int, RatingDistribution, DummyFeatureSource> Create(
IStarRatingRecommenderMapping <SplitInstanceSource <RecommenderDataset>, RatedUserItem, User, Item, int, DummyFeatureSource, Vector> mapping)
{
var inferNetRecommender = MatchboxRecommender.Create(mapping);
this.SetupMatchboxModel(inferNetRecommender.Settings);
return(inferNetRecommender);
}
public void RecommenderDatasetExample()
{
GenerateRecommenderDatasetFile("RatingsDataset.csv");
RecommenderDataset trainingDataset = RecommenderDataset.Load("RatingsDataset.csv");
var recommender = MatchboxRecommender.Create(Mappings.StarRatingRecommender);
recommender.Settings.Training.TraitCount = 5;
recommender.Settings.Training.IterationCount = 20;
recommender.Train(trainingDataset);
var recommendations = recommender.Recommend(new User("Person 1", Vector.FromArray(2.3)), 10);
CheckRecommendations(recommendations.Select(item => item.Id));
}
public void CsvMappingExample()
{
GenerateCsvFile("Ratings.csv");
var dataMapping = new CsvMapping();
var recommender = MatchboxRecommender.Create(dataMapping);
recommender.Settings.Training.TraitCount = 5;
recommender.Settings.Training.IterationCount = 20;
recommender.Train("Ratings.csv");
var recommendations = recommender.Recommend("Person 1", 10);
CheckRecommendations(recommendations);
}
static void Main(string[] args)
{
var dataMapping = new CsvMapping();
var recommender = MatchboxRecommender.Create(dataMapping);
recommender.Settings.Training.TraitCount = 5;
recommender.Settings.Training.IterationCount = 20;
recommender.Train("C:\\Python\\data\\ml-latest-small\\ratings.csv");
recommender.Save("C:\\Python\\data\\ml-latest-small\\TrainedModel.bin");
var recommends = recommender.Recommend("1", 10);
int rating = recommender.Predict("23", "6669");
rating = recommender.Predict("1", "31");
rating = recommender.Predict("2", "17");
}
public void BinaryRecommendationRegressionTest()
{
// TODO: We need a static class which implements mapping chainings
Rand.Restart(12347);
var positiveOnlyMapping = new PositiveOnlyDataMapping();
var negativeDataGeneratorMapping = positiveOnlyMapping.WithGeneratedNegativeData(1.0);
var recommender = MatchboxRecommender.Create(negativeDataGeneratorMapping);
recommender.Train(this.positiveOnlyDataset, this.featureProvider); // must not throw
foreach (var instance in positiveOnlyMapping.GetInstances(this.positiveOnlyDataset))
{
var user = positiveOnlyMapping.GetUser(this.positiveOnlyDataset, instance);
var item = positiveOnlyMapping.GetItem(this.positiveOnlyDataset, instance);
Assert.Equal(1, recommender.Predict(user, item, this.featureProvider));
}
}
GetRecommender(
IStarRatingRecommenderMapping <SplitInstanceSource <string>, RatingTriple, string, Movie, int, NoFeatureSource, Vector> dataMapping,
int traitCount)
{
// In this chapter Infer.NET Learner is used to make recommendations
// You can find its source code here:
// https://github.com/dotnet/infer/tree/master/src/Learners/Recommender
// This recommender has a model that matches the one described in the book.
// Its source code is situated here:
// https://github.com/dotnet/infer/blob/master/src/Learners/RecommenderModels/Models.cs
var recommender = MatchboxRecommender.Create(dataMapping);
if (traitCount == 0)
{
//workaround to make recommender work with 0 traits
var training = recommender.Settings.Training;
var countField = training.GetType().GetField("traitCount", BindingFlags.NonPublic | BindingFlags.Instance);
countField.SetValue(training, 0);
}
else
{
recommender.Settings.Training.TraitCount = traitCount;
}
// This value allows to normalize the sum of trait affinity variances to 1.0.
var traitVariance = 1.0 / Math.Sqrt(traitCount);
recommender.Settings.Training.IterationCount = IterationCount;
recommender.Settings.Training.BatchCount = 1;
recommender.Settings.Training.UseUserFeatures = false;
recommender.Settings.Training.UseItemFeatures = false;
recommender.Settings.Training.Advanced.AffinityNoiseVariance = 1.0;
recommender.Settings.Training.Advanced.UserThresholdNoiseVariance = EpsilonPriorVariance;
recommender.Settings.Training.Advanced.ItemTraitVariance = traitVariance;
recommender.Settings.Training.Advanced.ItemBiasVariance = 1.0;
recommender.Settings.Training.Advanced.UserTraitVariance = traitVariance;
recommender.Settings.Training.Advanced.UserBiasVariance = 1.0;
recommender.Settings.Prediction.SetPredictionLossFunction(LossFunction.ZeroOne);
return(recommender);
}
/// <summary>
/// Runs the module.
/// </summary>
/// <param name="args">The command line arguments for the module.</param>
/// <param name="usagePrefix">The prefix to print before the usage string.</param>
/// <returns>True if the run was successful, false otherwise.</returns>
public override bool Run(string[] args, string usagePrefix)
{
string trainingDatasetFile = string.Empty;
string trainedModelFile = string.Empty;
int traitCount = MatchboxRecommenderTrainingSettings.TraitCountDefault;
int iterationCount = MatchboxRecommenderTrainingSettings.IterationCountDefault;
int batchCount = MatchboxRecommenderTrainingSettings.BatchCountDefault;
bool useUserFeatures = MatchboxRecommenderTrainingSettings.UseUserFeaturesDefault;
bool useItemFeatures = MatchboxRecommenderTrainingSettings.UseItemFeaturesDefault;
var parser = new CommandLineParser();
parser.RegisterParameterHandler("--training-data", "FILE", "Training dataset", v => trainingDatasetFile = v, CommandLineParameterType.Required);
parser.RegisterParameterHandler("--trained-model", "FILE", "Trained model file", v => trainedModelFile = v, CommandLineParameterType.Required);
parser.RegisterParameterHandler("--traits", "NUM", "Number of traits (defaults to " + traitCount + ")", v => traitCount = v, CommandLineParameterType.Optional);
parser.RegisterParameterHandler("--iterations", "NUM", "Number of inference iterations (defaults to " + iterationCount + ")", v => iterationCount = v, CommandLineParameterType.Optional);
parser.RegisterParameterHandler("--batches", "NUM", "Number of batches to split the training data into (defaults to " + batchCount + ")", v => batchCount = v, CommandLineParameterType.Optional);
parser.RegisterParameterHandler("--use-user-features", "Use user features in the model (defaults to " + useUserFeatures + ")", () => useUserFeatures = true);
parser.RegisterParameterHandler("--use-item-features", "Use item features in the model (defaults to " + useItemFeatures + ")", () => useItemFeatures = true);
if (!parser.TryParse(args, usagePrefix))
{
return(false);
}
RecommenderDataset trainingDataset = RecommenderDataset.Load(trainingDatasetFile);
var recommender = MatchboxRecommender.Create(Mappings.StarRatingRecommender);
recommender.Settings.Training.TraitCount = traitCount;
recommender.Settings.Training.IterationCount = iterationCount;
recommender.Settings.Training.BatchCount = batchCount;
recommender.Settings.Training.UseUserFeatures = useUserFeatures;
recommender.Settings.Training.UseItemFeatures = useItemFeatures;
recommender.Train(trainingDataset);
recommender.Save(trainedModelFile);
return(true);
}
