transform

Type Members

1. trait ContinuousTransform extends Serializable

   

   A trait that describes a transformation for preprocessing continuous data.

   A trait that describes a transformation for preprocessing continuous data. The trait can be serialized and the implementation itself is responsible for storing the parameters for the transformation so that it can be used to modify new data examples.

   Created by cd on 14/05/2016.

2. case class IdentityTransform() extends ContinuousTransform with Product with Serializable

   

   The identity transform will return the input unchanged.

   The identity transform will return the input unchanged. Created by cd on 14/05/2016.

3. case class MinMaxNormalisation() extends ContinuousTransform with Product with Serializable

   

   min max normalisation will use columnwise minimum and columnwise maximum to normalise the data.

   min max normalisation will use columnwise minimum and columnwise maximum to normalise the data. It will store the minimum and maximum values for columns in order to be able to invert the transform.

4. case class StandardisedNormalisation() extends ContinuousTransform with Product with Serializable

   

   The data will be transformed using the empircal mean and approximate variance to create a spherical result.

   The data will be transformed using the empircal mean and approximate variance to create a spherical result.

   The standardised normalisation will store the parameters \hat{mu} and \hat{sigma}

   Created by cd on 14/05/2016.