Skip to content
forked from pa-m/sklearn

attempt to port bits of sklearn to golang go

License

Notifications You must be signed in to change notification settings

sbinet-gonum/sklearn

 
 

Repository files navigation

sklearn

Very partial port of scikit-learn to go

Build Status Code Coverage Go Report Card GoDoc

for now, ported only some estimators including

You'll also find

  • some metrics MeanSquaredError,MeanAbsoluteError,R2Score,AccuracyScore, ...
  • some preprocessing MinMaxScaler,StandardScaler,OneHotEncoder,PolynomialFeatures
  • Pipeline and MakePipeline
  • some interpolation stuff like in scipy.interpolate: interp1d,interp2d,CubicSpline
  • all estimators can use following
    • solvers: sgd,adagrad,rmsprop,adadelta,adam + all gonum/optimize methods
    • loss functions: square,cross-entropy
    • activation functions: identity,logistic,tanh,relu

All of this is

  • a personal project to get a deeper understanding of how all of this magic works
  • a recent work still in progress, subject to refactoring, so interfaces may change, especially args to NewXXX
  • processed with gofmt, golint, go vet
  • unit tested but coverage should reach 90%
  • underdocumented but python sklearn documentation should be sufficient

Many thanks to gonum and scikit-learn contributors

PRs are welcome

Best regards

About

attempt to port bits of sklearn to golang go

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages

  • Go 100.0%