Safe Haskell | None |
---|---|
Language | Haskell2010 |
This library provides generic machinery (via GHC.Generics and `generics-sop`) to encode values of some algebraic type as points in a vector space.
Processing datasets that have one or more categorical variables (which in other words are values of a sum type) typically requires a series of boilerplate transformations, and the encodeOneHot
function provided here does precisely that.
Internals
This library makes use of generic programming to analyze both values and types (see the internal Data.Record.Encode.Generics module).
Initially, it was relying on Template Haskell to analyze types, using the the instance generation machinery explained here: https://markkarpov.com/tutorial/th.html#example-1-instance-generation
One-hot encoding
encodeOneHot :: forall a. G a => a -> OneHot Source #
Computes the one-hot encoding of a value of a sum type.
The type of the input value must be an instance of Generic
(from GHC.Generics) and of Generic
(from the `generics-sop` library).
>>>
:set -XDeriveGeneric
>>>
import qualified GHC.Generics as G
>>>
import qualified Generics.SOP as SOP
>>>
import Data.Record.Encode
>>>
data X = A | B | C deriving (G.Generic)
>>>
instance SOP.Generic X
>>>
encodeOneHot B
OH {oDim = 3, oIx = 1}
Types and Utilities
A one-hot encoding is a d-dimensional vector having a single component equal to 1 and all others equal to 0.