transformTS: Time series format transformations

In jmaspons/MLTools: Pipes and tools for machine learning

Time series format transformations

Description

For data in wide format, the name of time varying columns follow the pattern var_time.

Usage

longToWide.ts(d, timevar, idCols = NULL)

wideToLong.ts(d, timevar, vars, idCols = NULL, regex_time = ".+")

wideTo3Darray.ts(d, vars, idCols = NULL)

longTo3Darray.ts(d, timevar, idCols = NULL)

Arguments

d	a data.frame or matrix.
timevar	column name for the time.
idCols	ids or no time varying columns. Works with column names and indexes.
vars	time varying variables in the long format (ie. without the _time part in wide format).
regex_time	a regular expression for the time part of the variable names for data in wide format.

Value

the same data in a wide [var_time columns], long [time + vars columns] or a 3D array [samples, time, vars] format.

jmaspons/MLTools documentation built on Jan. 27, 2024, 4:31 a.m.