dataframe-learn-2.4.1.0: Interpretable, expression-returning machine learning for the dataframe ecosystem.
Safe HaskellNone
LanguageHaskell2010

DataFrame.LinearModel.Logistic

Description

Logistic regression: binary and one-vs-rest multiclass over the FISTA solver. fit trains a LogisticModel; predict is the arg-max class decision. Per-class margins and (normalized) probabilities stay available as named auxiliary expressions.

Synopsis

Documentation

module DataFrame.Model

newtype LogisticConfig Source #

Hyperparameters for logistic regression (the FISTA solver config).

Constructors

LogisticConfig 

Fields

Instances

Instances details
Show LogisticConfig Source # 
Instance details

Defined in DataFrame.LinearModel.Logistic

Methods

showsPrec :: Int -> LogisticConfig -> ShowS #

show :: LogisticConfig -> String #

showList :: [LogisticConfig] -> ShowS #

Eq LogisticConfig Source # 
Instance details

Defined in DataFrame.LinearModel.Logistic

Methods

(==) :: LogisticConfig -> LogisticConfig -> Bool #

(/=) :: LogisticConfig -> LogisticConfig -> Bool #

(Columnable a, Ord a) => SegmentFit LogisticConfig a Source #

Logistic segments support independent fitting (lambda = 0) only, for now.

Instance details

Defined in DataFrame.Segmented

Methods

fitSegments :: LogisticConfig -> Double -> Expr a -> [DataFrame] -> [ModelOf LogisticConfig (Expr a)] Source #

(Columnable a, Ord a) => Fit LogisticConfig (Expr a) Source # 
Instance details

Defined in DataFrame.LinearModel.Logistic

Associated Types

type ModelOf LogisticConfig (Expr a) 
Instance details

Defined in DataFrame.LinearModel.Logistic

type ModelOf LogisticConfig (Expr a) = LogisticModel a
type FrameReq LogisticConfig (Expr a) 
Instance details

Defined in DataFrame.LinearModel.Logistic

type FrameReq LogisticConfig (Expr a) = 'AnyFrame

Methods

fit :: LogisticConfig -> Expr a -> FrameFor (Expr a) -> FitResult (FrameFor (Expr a)) (ModelOf LogisticConfig (Expr a)) Source #

type FrameReq LogisticConfig (Expr a) Source # 
Instance details

Defined in DataFrame.LinearModel.Logistic

type FrameReq LogisticConfig (Expr a) = 'AnyFrame
type ModelOf LogisticConfig (Expr a) Source # 
Instance details

Defined in DataFrame.LinearModel.Logistic

type ModelOf LogisticConfig (Expr a) = LogisticModel a

defaultLogisticConfig :: LogisticConfig Source #

data LogisticModel a Source #

A fitted (one-vs-rest) logistic model: parallel vectors of class labels and their binary sub-models. lgModels carries sklearn's per-class coef_.

Constructors

LogisticModel 

Fields

Instances

Instances details
Show a => Show (LogisticModel a) Source # 
Instance details

Defined in DataFrame.LinearModel.Logistic

Methods

showsPrec :: Int -> LogisticModel a -> ShowS #

show :: LogisticModel a -> String #

showList :: [LogisticModel a] -> ShowS #

(Columnable a, Ord a) => Predict (LogisticModel a) Source # 
Instance details

Defined in DataFrame.LinearModel.Logistic

Associated Types

type Prediction (LogisticModel a) 
Instance details

Defined in DataFrame.LinearModel.Logistic

type Prediction (LogisticModel a) = Expr a

Methods

predict :: LogisticModel a -> Prediction (LogisticModel a) Source #

Eq a => Eq (LogisticModel a) Source # 
Instance details

Defined in DataFrame.LinearModel.Logistic

Methods

(==) :: LogisticModel a -> LogisticModel a -> Bool #

(/=) :: LogisticModel a -> LogisticModel a -> Bool #

type Prediction (LogisticModel a) Source # 
Instance details

Defined in DataFrame.LinearModel.Logistic

type Prediction (LogisticModel a) = Expr a

logisticMarginExprs :: (Columnable a, Ord a) => LogisticModel a -> Map a (Expr Double) Source #

The raw margin Expr for each class.

logisticProbExprs :: (Columnable a, Ord a) => LogisticModel a -> Map a (Expr Double) Source #

Per-class probability expressions: 1 / (1 + exp(-margin)).