Search Results for

    Show / Hide Table of Contents

    Interface IMatchboxRecommenderMapping<TInstanceSource, TFeatureSource>

    A mapping used by the implementations of IMatchboxRecommender<TInstanceSource, TUser, TItem, TRatingDistribution, TFeatureSource> interface to convert user data to an appropriate format.

    Namespace: Microsoft.ML.Probabilistic.Learners.Mappings
    Assembly: Microsoft.ML.Probabilistic.Learners.Recommender.dll
    Syntax
    public interface IMatchboxRecommenderMapping<in TInstanceSource, in TFeatureSource>
    Type Parameters
    Name Description
    TInstanceSource

    The type of a source of instances.

    TFeatureSource

    The type of a source of features.

    Remarks

    Feature-related methods in this interface are only called when user of item features are being used. During training, features for all entities are required, and hence the methods called are the GetAllFeature ones. During prediction, features are requested on per-entity basis, and hence the GetSingleFeature methods are called. Note that entity features are expected not to change between training and prediction, and therefore only features of cold users or items will be requested during prediction.

    Methods

    GetAllItemNonZeroFeatureIndices(TFeatureSource)

    Gets non-zero feature indices for all items present in a given feature source.

    Declaration
    IReadOnlyList<IReadOnlyList<int>> GetAllItemNonZeroFeatureIndices(TFeatureSource featureSource)
    Parameters
    Type Name Description
    TFeatureSource featureSource

    The source to obtain feature indices from.

    Returns
    Type Description
    IReadOnlyList<IReadOnlyList<Int32>>

    An array of non-zero item feature index arrays where the outer array is indexed by item id

    Remarks

    This function will be called during training if the item feature support is enabled.

    GetAllItemNonZeroFeatureValues(TFeatureSource)

    Gets non-zero feature values for all items present in a given feature source.

    Declaration
    IReadOnlyList<IReadOnlyList<double>> GetAllItemNonZeroFeatureValues(TFeatureSource featureSource)
    Parameters
    Type Name Description
    TFeatureSource featureSource

    The source to obtain features from.

    Returns
    Type Description
    IReadOnlyList<IReadOnlyList<Double>>

    An array of non-zero item feature arrays where the outer array is indexed by item id

    Remarks

    This function will be called during training if the item feature support is enabled.

    GetAllUserNonZeroFeatureIndices(TFeatureSource)

    Gets non-zero feature indices for all users present in a given feature source.

    Declaration
    IReadOnlyList<IReadOnlyList<int>> GetAllUserNonZeroFeatureIndices(TFeatureSource featureSource)
    Parameters
    Type Name Description
    TFeatureSource featureSource

    The source to obtain feature indices from.

    Returns
    Type Description
    IReadOnlyList<IReadOnlyList<Int32>>

    An array of non-zero user feature index arrays where the outer array is indexed by user id.

    Remarks

    This function will be called during training if the user feature support is enabled.

    GetAllUserNonZeroFeatureValues(TFeatureSource)

    Gets non-zero feature values for all users present in a given feature source.

    Declaration
    IReadOnlyList<IReadOnlyList<double>> GetAllUserNonZeroFeatureValues(TFeatureSource featureSource)
    Parameters
    Type Name Description
    TFeatureSource featureSource

    The source to obtain features from.

    Returns
    Type Description
    IReadOnlyList<IReadOnlyList<Double>>

    An array of non-zero user feature arrays where the outer array is indexed by user id.

    Remarks

    This function will be called during training if the user feature support is enabled.

    GetItemCount(TInstanceSource)

    Gets the number of items from a given instance source.

    Declaration
    int GetItemCount(TInstanceSource instanceSource)
    Parameters
    Type Name Description
    TInstanceSource instanceSource

    The source of instances to get number of items from.

    Returns
    Type Description
    Int32

    The number of items.

    GetItemFeatureCount(TFeatureSource)

    Gets the number of item features.

    Declaration
    int GetItemFeatureCount(TFeatureSource featureSource)
    Parameters
    Type Name Description
    TFeatureSource featureSource

    The source to obtain features from.

    Returns
    Type Description
    Int32

    The number of item features.

    GetItemIds(TInstanceSource, Int32)

    Gets the list of item identifiers from a given instance source.

    Declaration
    IReadOnlyList<int> GetItemIds(TInstanceSource instanceSource, int batchNumber = 0)
    Parameters
    Type Name Description
    TInstanceSource instanceSource

    The source of instances to get the item identifiers from.

    Int32 batchNumber

    The number of the current batch (used only if the data is divided into batches).

    Returns
    Type Description
    IReadOnlyList<Int32>

    The list of item identifiers.

    Remarks

    Note that this method can be called concurrently from multiple threads, so its implementation must be thread-safe.

    GetRatingCount(TInstanceSource)

    Gets the number of star ratings. This is equal to one plus the difference between the maximum and the minimum rating.

    Declaration
    int GetRatingCount(TInstanceSource instanceSource)
    Parameters
    Type Name Description
    TInstanceSource instanceSource

    The source of instances to get number of items from.

    Returns
    Type Description
    Int32

    The number of ratings.

    GetRatings(TInstanceSource, Int32)

    Gets the list of ratings from a given instance source.

    Declaration
    IReadOnlyList<int> GetRatings(TInstanceSource instanceSource, int batchNumber = 0)
    Parameters
    Type Name Description
    TInstanceSource instanceSource

    The source of instances to get the ratings from.

    Int32 batchNumber

    The number of the current batch (used only if the data is divided into batches).

    Returns
    Type Description
    IReadOnlyList<Int32>

    The list of ratings

    Remarks

    Note that this method can be called concurrently from multiple threads, so its implementation must be thread-safe.

    GetSingleItemNonZeroFeatureIndices(TFeatureSource, Int32)

    Gets non-zero feature indices for a given item.

    Declaration
    IReadOnlyList<int> GetSingleItemNonZeroFeatureIndices(TFeatureSource featureSource, int itemId)
    Parameters
    Type Name Description
    TFeatureSource featureSource

    The source to obtain feature indices from.

    Int32 itemId

    The item identifier.

    Returns
    Type Description
    IReadOnlyList<Int32>

    Non-zero feature values for the item.

    Remarks

    This function will be called during prediction for cold items if the item feature support is enabled.

    GetSingleItemNonZeroFeatureValues(TFeatureSource, Int32)

    Gets non-zero feature values for a given item.

    Declaration
    IReadOnlyList<double> GetSingleItemNonZeroFeatureValues(TFeatureSource featureSource, int itemId)
    Parameters
    Type Name Description
    TFeatureSource featureSource

    The source to obtain features from.

    Int32 itemId

    The item identifier.

    Returns
    Type Description
    IReadOnlyList<Double>

    Non-zero feature values for the item.

    Remarks

    This function will be called during prediction for cold items if the item feature support is enabled.

    GetSingleUserNonZeroFeatureIndices(TFeatureSource, Int32)

    Gets non-zero feature indices for a given user.

    Declaration
    IReadOnlyList<int> GetSingleUserNonZeroFeatureIndices(TFeatureSource featureSource, int userId)
    Parameters
    Type Name Description
    TFeatureSource featureSource

    The source to obtain feature indices from.

    Int32 userId

    The user identifier.

    Returns
    Type Description
    IReadOnlyList<Int32>

    Non-zero feature indices for the user.

    Remarks

    This function will be called during prediction for cold users if the user feature support is enabled.

    GetSingleUserNonZeroFeatureValues(TFeatureSource, Int32)

    Gets non-zero feature values for a given user.

    Declaration
    IReadOnlyList<double> GetSingleUserNonZeroFeatureValues(TFeatureSource featureSource, int userId)
    Parameters
    Type Name Description
    TFeatureSource featureSource

    The source to obtain features from.

    Int32 userId

    The user identifier.

    Returns
    Type Description
    IReadOnlyList<Double>

    Non-zero feature values for the user.

    Remarks

    This function will be called during prediction for cold users if the user feature support is enabled.

    GetUserCount(TInstanceSource)

    Gets the number of users from a given instance source.

    Declaration
    int GetUserCount(TInstanceSource instanceSource)
    Parameters
    Type Name Description
    TInstanceSource instanceSource

    The source of instances to get number of users from.

    Returns
    Type Description
    Int32

    The number of users.

    GetUserFeatureCount(TFeatureSource)

    Gets the number of user features.

    Declaration
    int GetUserFeatureCount(TFeatureSource featureSource)
    Parameters
    Type Name Description
    TFeatureSource featureSource

    The source to obtain features from.

    Returns
    Type Description
    Int32

    The number of user features.

    GetUserIds(TInstanceSource, Int32)

    Gets the list of user identifiers from a given instance source.

    Declaration
    IReadOnlyList<int> GetUserIds(TInstanceSource instanceSource, int batchNumber = 0)
    Parameters
    Type Name Description
    TInstanceSource instanceSource

    The source of instances to get the user identifiers from.

    Int32 batchNumber

    The number of the current batch (used only if the data is divided into batches).

    Returns
    Type Description
    IReadOnlyList<Int32>

    The list of user identifiers.

    Remarks

    Note that this method can be called concurrently from multiple threads, so its implementation must be thread-safe.

    In This Article
    Back to top Copyright © .NET Foundation. All rights reserved.