共用方式為


FieldAwareFactorizationMachineModelParameters 類別

定義

public sealed class FieldAwareFactorizationMachineModelParameters : Microsoft.ML.Trainers.ModelParametersBase<float>
type FieldAwareFactorizationMachineModelParameters = class
    inherit ModelParametersBase<single>
Public NotInheritable Class FieldAwareFactorizationMachineModelParameters
Inherits ModelParametersBase(Of Single)
繼承
FieldAwareFactorizationMachineModelParameters

屬性

FeatureCount

取得功能數目。 這是檔中的符號 nhttps://github.com/wschin/fast-ffm/blob/master/fast-ffm.pdf

FieldCount

取得欄位數目。 這是檔中的符號 mhttps://github.com/wschin/fast-ffm/blob/master/fast-ffm.pdf

LatentDimension

取得延遲維度。 這是檔中的 v_{j, f} 長度: https://github.com/wschin/fast-ffm/blob/master/fast-ffm.pdf

方法

GetLatentWeights()

每個功能的延遲標記法。 請注意,其中一個特徵可能會有 FieldCount 晚期向量,而每個延遲向量都包含 LatentDimension 值。 在 f-th 欄位中,第 j 個特徵在檔中 https://github.com/wschin/fast-ffm/blob/master/fast-ffm.pdf 的延遲向量 v_{j, f} 從 latentWeights[j * fieldCount * latentDim + f * latentDim 開始。 v_{j, f} 中的 k 元素是 latentWeights[j * fieldCount * latentDim + f * latentDim + k]。 傳回值的大小為 featureCount x fieldCount x latentDim。

GetLinearWeights()

特徵的線性係數。 這是檔中的符號 whttps://github.com/wschin/fast-ffm/blob/master/fast-ffm.pdf

明確介面實作

ICanSaveModel.Save(ModelSaveContext)

FieldAwareFactorizationMachineTrainer 模型參數。

(繼承來源 ModelParametersBase<TOutput>)

適用於