Safe Haskell	None
Language	Haskell2010

DataFrame.PCA

Description

Principal component analysis via the symmetric Jacobi eigensolver on the covariance of the (optionally standardized) feature columns. fit trains a PCAModel (components + explained variance); the projection is exposed as pcaExprs / pcaTransform (PCA is a transformer, so it has no Predict).

Synopsis

module DataFrame.Model
data NComponents
- = NComp !Int
- | VarianceCovered !Double
data PCAConfig = PCAConfig {
- pcaNComponents :: !NComponents
- pcaStandardize :: !Bool
}
defaultPCAConfig :: PCAConfig
data PCAModel = PCAModel {
- pcaComponents :: !(Vector (Vector Double))
- pcaExplainedVariance :: !(Vector Double)
- pcaExplainedVarianceRatio :: !(Vector Double)
- pcaMean :: !(Vector Double)
- pcaScale :: !(Maybe (Vector Double))
- pcaFeatureNames :: !(Vector Text)
}
pcaExprs :: PCAModel -> [(Text, Expr Double)]
pcaTransform :: PCAModel -> Transform

Documentation

module DataFrame.Model

data NComponents Source #

How many components to keep.

Constructors

NComp !Int
VarianceCovered !Double

Instances

Instances details

Show NComponents Source #
Instance details Defined in DataFrame.PCA Methods showsPrec :: Int -> NComponents -> ShowS # show :: NComponents -> String # showList :: [NComponents] -> ShowS #
Eq NComponents Source #
Instance details Defined in DataFrame.PCA Methods (==) :: NComponents -> NComponents -> Bool # (/=) :: NComponents -> NComponents -> Bool #

data PCAConfig Source #

Constructors

PCAConfig
Fields pcaNComponents :: !NComponents pcaStandardize :: !Bool

Instances

Instances details

Show PCAConfig Source #

Instance details

Defined in DataFrame.PCA

Methods

showsPrec :: Int -> PCAConfig -> ShowS #

show :: PCAConfig -> String #

showList :: [PCAConfig] -> ShowS #

Eq PCAConfig Source #

Instance details

Defined in DataFrame.PCA

Methods

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

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

Fit PCAConfig [Expr Double] Source #

Instance details

Defined in DataFrame.PCA

Associated Types

type ModelOf PCAConfig [Expr Double]
Instance details Defined in DataFrame.PCA type ModelOf PCAConfig [Expr Double] = PCAModel
type FrameReq PCAConfig [Expr Double]
Instance details Defined in DataFrame.PCA type FrameReq PCAConfig [Expr Double] = 'AnyFrame

Methods

fit :: PCAConfig -> [Expr Double] -> FrameFor [Expr Double] -> FitResult (FrameFor [Expr Double]) (ModelOf PCAConfig [Expr Double]) Source #

type FrameReq PCAConfig [Expr Double] Source #

Instance details

Defined in DataFrame.PCA

type FrameReq PCAConfig [Expr Double] = 'AnyFrame

type ModelOf PCAConfig [Expr Double] Source #

Instance details

Defined in DataFrame.PCA

type ModelOf PCAConfig [Expr Double] = PCAModel

defaultPCAConfig :: PCAConfig Source #

data PCAModel Source #

A fitted PCA. pcaComponents are sklearn's components_ (row i is the i-th loading vector); pcaScale is Just the per-column std when standardizing.

Constructors

PCAModel
Fields pcaComponents :: !(Vector (Vector Double)) pcaExplainedVariance :: !(Vector Double) pcaExplainedVarianceRatio :: !(Vector Double) pcaMean :: !(Vector Double) pcaScale :: !(Maybe (Vector Double)) pcaFeatureNames :: !(Vector Text)

Instances

Instances details

Show PCAModel Source #
Instance details Defined in DataFrame.PCA Methods showsPrec :: Int -> PCAModel -> ShowS # show :: PCAModel -> String # showList :: [PCAModel] -> ShowS #
Eq PCAModel Source #
Instance details Defined in DataFrame.PCA Methods (==) :: PCAModel -> PCAModel -> Bool # (/=) :: PCAModel -> PCAModel -> Bool #

pcaExprs :: PCAModel -> [(Text, Expr Double)] Source #

Per-component projection expressions, named pc1, pc2, …

pcaTransform :: PCAModel -> Transform Source #

The PCA projection as a composable fitted Transform.