Re-Scale Vectors and Time-Series Features

Provides standardized access to a range of re-scaling methods for numerical vectors and time-series features calculated within the 'theft' ecosystem.

normaliseR

Re-Scale Vectors and Time-Series Features

Installation

You can install the development version of normaliseR from GitHub using the following:

devtools::install_github("hendersontrent/normaliseR")

General purpose

normaliseR is a software package for R for rescaling numerical vectors or feature_calculations objects produced by the theft R package for computing time-series features.

Putting calculated feature vectors on an equal scale is crucial for any statistical or machine learning model as variables with high variance can adversely impact the model’s capacity to fit the data appropriately, learn appropriate weight values, or minimise a loss function. normaliseR includes function normalise (or normalize) to rescale either a whole feature_calculations object, or a single vector of values. The following normalisation methods are currently offered:

• z-score—"zScore"
• Sigmoid—"Sigmoid"
• Outlier-robust Sigmoid (credit to Ben Fulcher for creating the original MATLAB version) – "RobustSigmoid"
• Min-max—"MinMax"
• Maximum absolute—"MaxAbs"