Documentation ¶
Overview ¶
Package hyperopt implements SMBO/TPE hyper-parameter optimization for ML models
Many thanks to Masashi SHIBATA for his excellent work on goptuna I used github.com/c-bata/goptuna as a reference implementation for the paper 'Algorithms for Hyper-Parameter Optimization' https://papers.nips.cc/paper/4443-algorithms-for-hyper-parameter-optimization.pdf
TPE sampler mostly derived from goptuna.
Index ¶
Constants ¶
This section is empty.
Variables ¶
This section is empty.
Functions ¶
Types ¶
type BestParams ¶
BestParams is a result of Hyper-parameters Optimization
type IntRange ¶
type IntRange [2]int
IntRange is a close integer range specified by min and max values [min,max]
type LogIntRange ¶
type LogIntRange [2]int
LogRange is a close logarithmic integer range specified by min and max values [min,max]
type LogRange ¶
type LogRange [2]float64
LogRange is a open float logarithmic range specified by min and max values (min,max)
type Range ¶
type Range [2]float64
Range is a open float range specified by min and max values (min,max)
type Space ¶
type Space struct { Source tables.AnyData // dataset source Features []string // dataset features Label string // dataset label Seed int // random seed Kfold int // count of dataset folds Iterations int // model fitting iterations Metrics model.Metrics // model evaluation metrics Score model.Score // function to calculate score of train/test metrics // the model generation function ModelFunc func(Params) model.HungryModel // hyper-parameters variance Variance Variance }
Space is a definition of hyper-parameters optimization space
func (Space) LuckyOptimize ¶
func (ss Space) LuckyOptimize(trails int) BestParams