MatrixFactorizationTrainer.Options.LossFunction Field

  • Reference

Definition

Namespace:
Microsoft.ML.Trainers
Assembly:
Microsoft.ML.Recommender.dll
Package:
Microsoft.ML.Recommender v0.21.1

Loss function minimized for finding factor matrices.

public Microsoft.ML.Trainers.MatrixFactorizationTrainer.LossFunctionType LossFunction;
val mutable LossFunction : Microsoft.ML.Trainers.MatrixFactorizationTrainer.LossFunctionType
Public LossFunction As MatrixFactorizationTrainer.LossFunctionType

Field Value

MatrixFactorizationTrainer.LossFunctionType

Remarks

Two values are allowed, SquareLossRegression or SquareLossOneClass. The SquareLossRegression means traditional collaborative filtering problem with squared loss. The SquareLossOneClass triggers one-class matrix factorization for implicit-feedback recommendation problem.

Applies to