Property | Type | Description |
---|
Method | Description | |
---|---|---|
RecommendationBuildParameters ( ) : System.Linq |
Initializes a new instance of the RecommendationBuildParameters class.
|
|
RecommendationBuildParameters ( int numberOfModelIterations = default(int?), int numberOfModelDimensions = default(int?), int itemCutOffLowerBound = default(int?), int itemCutOffUpperBound = default(int?), int userCutOffLowerBound = default(int?), int userCutOffUpperBound = default(int?), bool enableModelingInsights = default(bool?), string splitterStrategy = default(string), RandomSplitterParameters randomSplitterParameters = default(RandomSplitterParameters), |
Initializes a new instance of the RecommendationBuildParameters class.
|
public RecommendationBuildParameters ( ) : System.Linq | ||
return | System.Linq |
public RecommendationBuildParameters ( int numberOfModelIterations = default(int?), int numberOfModelDimensions = default(int?), int itemCutOffLowerBound = default(int?), int itemCutOffUpperBound = default(int?), int userCutOffLowerBound = default(int?), int userCutOffUpperBound = default(int?), bool enableModelingInsights = default(bool?), string splitterStrategy = default(string), RandomSplitterParameters randomSplitterParameters = default(RandomSplitterParameters), |
||
numberOfModelIterations | int | The number of iterations the /// model performs. /// The higher the number, the better accuracy, but /// compute time will be higher. |
numberOfModelDimensions | int | The number of dimensions /// relates to the number of 'features' the model will try to find /// within your data. /// Increasing the number of dimensions will allow better /// fine-tuning of the results into smaller clusters. /// However, too many dimensions will prevent the model /// from finding correlations between items. |
itemCutOffLowerBound | int | Defines the item lower bound /// for usage condenser. |
itemCutOffUpperBound | int | Defines the item upper bound /// for usage condenser. |
userCutOffLowerBound | int | Defines the user lower bound /// for usage condenser. |
userCutOffUpperBound | int | Defines the user upper bound /// for usage condenser. |
enableModelingInsights | bool | Enable or disable metrics /// computation for the model. |
splitterStrategy | string | Defines the splitter strategy to be /// used by the build. /// RandomSplitter splits the usage data in train and test /// sets based on the given /// randomSplitterParameters value. /// LastEventSplitter splits the usage data in train and /// test sets based on the last /// transaction for a each user. |
randomSplitterParameters | RandomSplitterParameters | Specifies the parameters to /// be used for random splitter. |
dateSplitterParameters | Specifies the parameters to /// be used for date splitter. | |
popularItemBenchmarkWindow | int | Specifies the parameters /// to be used for computing popular items for modeling insights. (in /// number of days) |
useFeaturesInModel | bool | Indicates if features can be used /// in order to enhance the recommendation model. |
modelingFeatureList | string | Comma-separated list of feature /// names to be used during build. |
allowColdItemPlacement | bool | Indicates if the /// recommendation should also push cold items via feature /// similarity. |
enableFeatureCorrelation | bool | Indicates if features can /// be used in reasoning. |
reasoningFeatureList | string | Comma-separated list of feature /// names to be used for reasoning sentences (e.g. recommendation /// explanations). |
enableU2I | bool | Allow the personalized recommendation /// a.k.a. U2I (user to item recommendations). |
return | System.Linq |