MultivariateLinearRegressor

Declaration

struct MultivariateLinearRegressor<Scalar> where Scalar : MLShapedArrayScalar, Scalar : BinaryFloatingPoint

Overview

Unlike a LinearRegressor, a MultivariateLinearRegressor supports shaped array outputs with any number of elements. It also provides a wider range of training options better suited for large multi-dimensional regression.

Topics

Creating a regressor

• init(configuration:)

Getting the configuration

• configuration

Fitting

• fitted(to:eventHandler:)
• fitted(to:validateOn:eventHandler:)

Fitting Progressively

• makeTransformer()
• update(_:with:eventHandler:)
• update(_:with:)

Encoding and decoding

• encodeWithOptimizer(_:to:)
• decodeWithOptimizer(from:)

Supporting types

• MultivariateLinearRegressor.Model
• MultivariateLinearRegressor.Annotation
• MultivariateLinearRegressor.Configuration
• MultivariateLinearRegressor.Feature
• MultivariateLinearRegressor.Transformer

Default Implementations

• UpdatableSupervisedEstimator Implementations

Relationships

Conforms To

• Copyable
• Escapable
• Sendable
• SendableMetatype
• SupervisedEstimator
• UpdatableSupervisedEstimator

See Also

Regressors

• Regressor
• LinearRegressor
• LinearRegressorModel
• MultivariateLinearRegressorConfiguration
• MultivariateLinearRegressor.Model
• FullyConnectedNetworkRegressor
• FullyConnectedNetworkRegressorModel
• BoostedTreeRegressor
• TreeRegressorModel
• OptimizationStrategy