Copyright	(c) Fabricio Olivetti 2021 - 2024
License	BSD3
Maintainer	fabricio.olivetti@gmail.com
Stability	experimental
Portability	ConstraintKinds
Safe Haskell	None
Language	Haskell2010

Algorithm.SRTree.ModelSelection

Description

Helper functions for model selection criteria

Synopsis

bic :: EvaluatedTree -> Double
aic :: EvaluatedTree -> Double
evidence :: EvaluatedTree -> Double
fractionalBayesFactor :: EvaluatedTree -> Double
mdl :: EvaluatedTree -> Double
mdlLatt :: EvaluatedTree -> Double
mdlFreq :: EvaluatedTree -> Double
logFunctional :: Fix SRTree -> Double
logFunctionalFreq :: Fix SRTree -> Double
data ModelEval
- = RMSE
- | R2
- | AIC
- | BIC
- | Evidence
- | FBF
- | MDL
- | MDLLatt
- | MDLFreq
- | EvalLoss Loss
module Algorithm.SRTree.Compile

Documentation

bic :: EvaluatedTree -> Double Source #

Bayesian information criterion

aic :: EvaluatedTree -> Double Source #

Akaike information criterion

evidence :: EvaluatedTree -> Double Source #

Evidence

fractionalBayesFactor :: EvaluatedTree -> Double Source #

mdl :: EvaluatedTree -> Double Source #

MDL as described in Bartlett, Deaglan J., Harry Desmond, and Pedro G. Ferreira. "Exhaustive symbolic regression." IEEE Transactions on Evolutionary Computation (2023).

mdlLatt :: EvaluatedTree -> Double Source #

MDL Lattice as described in Bartlett, Deaglan, Harry Desmond, and Pedro Ferreira. "Priors for symbolic regression." Proceedings of the Companion Conference on Genetic and Evolutionary Computation. 2023.

mdlFreq :: EvaluatedTree -> Double Source #

same as mdl but weighting the functional structure by frequency calculated using a wiki information of physics and engineering functions

logFunctional :: Fix SRTree -> Double Source #

logFunctionalFreq :: Fix SRTree -> Double Source #

data ModelEval Source #

The possible metrics used to evaluate/select a fitted model, ranging from plain loss functions (EvalLoss, wrapping any Loss -- including a distribution's negative log-likelihood via EvalLoss (NLL dist)) to the error metrics and model-selection criteria already provided by this module (RMSE, R2, AIC, BIC, Evidence, FBF, MDL, MDLLatt, MDLFreq).

Constructors

RMSE
R2
AIC
BIC
Evidence
FBF
MDL
MDLLatt
MDLFreq
EvalLoss Loss

Instances

Instances details

Eq ModelEval Source #
Instance details Defined in Algorithm.SRTree.ModelSelection Methods (==) :: ModelEval -> ModelEval -> Bool # (/=) :: ModelEval -> ModelEval -> Bool #
Bounded ModelEval Source #
Instance details Defined in Algorithm.SRTree.ModelSelection Methods minBound :: ModelEval # maxBound :: ModelEval #
Enum ModelEval Source #
Instance details Defined in Algorithm.SRTree.ModelSelection Methods succ :: ModelEval -> ModelEval # pred :: ModelEval -> ModelEval # toEnum :: Int -> ModelEval # fromEnum :: ModelEval -> Int # enumFrom :: ModelEval -> [ModelEval] # enumFromThen :: ModelEval -> ModelEval -> [ModelEval] # enumFromTo :: ModelEval -> ModelEval -> [ModelEval] # enumFromThenTo :: ModelEval -> ModelEval -> ModelEval -> [ModelEval] #
Read ModelEval Source #
Instance details Defined in Algorithm.SRTree.ModelSelection Methods readsPrec :: Int -> ReadS ModelEval # readList :: ReadS [ModelEval] # readPrec :: ReadPrec ModelEval # readListPrec :: ReadPrec [ModelEval] #
Show ModelEval Source #
Instance details Defined in Algorithm.SRTree.ModelSelection Methods showsPrec :: Int -> ModelEval -> ShowS # show :: ModelEval -> String # showList :: [ModelEval] -> ShowS #

module Algorithm.SRTree.Compile