foregen icon indicating copy to clipboard operation
foregen copied to clipboard

(fo)recast (re)gressor (gen)erator

trafficstars

foregen

(fo)recast (re)gressor (gen)erator The motivation for this regressor generator is to be able to easily create regressors for use in time series regression and forecasting. It makes a data.frame that can be input into the xreg argument as seen in the forecast package. It also has two other functions to help in creating regressors from a table of events.

Installation

You can also install the development version from Github

# install.packages("devtools")
library(devtools)
install_github("republicwireless-open/foregen") 

Usage

library(foregen)

regs_table <- data.frame(event=c("christmas", "christmas", "promotion"),
                         start=c("2014-12-25", "2015-12-25", "2015-11-27"),
                         end=c("2014-12-25", "2015-12-25", "2015-11-30"),
                         time_function= c("none","none","inverse"),
                         stringsAsFactors=FALSE)

events_to_regressors(table_to_events(regs_table),
                     seq(as.Date("2015-11-26"), as.Date("2016-01-01"), by = 1))

           christmas promotion
2015-11-26         0 0.0000000
2015-11-27         0 1.0000000
2015-11-28         0 0.5000000
2015-11-29         0 0.3333333
2015-11-30         0 0.2500000
2015-12-01         0 0.0000000
2015-12-02         0 0.0000000
2015-12-03         0 0.0000000
2015-12-04         0 0.0000000
2015-12-05         0 0.0000000
2015-12-06         0 0.0000000
2015-12-07         0 0.0000000
2015-12-08         0 0.0000000
2015-12-09         0 0.0000000
2015-12-10         0 0.0000000
2015-12-11         0 0.0000000
2015-12-12         0 0.0000000
2015-12-13         0 0.0000000
2015-12-14         0 0.0000000
2015-12-15         0 0.0000000
2015-12-16         0 0.0000000
2015-12-17         0 0.0000000
2015-12-18         0 0.0000000
2015-12-19         0 0.0000000
2015-12-20         0 0.0000000
2015-12-21         0 0.0000000
2015-12-22         0 0.0000000
2015-12-23         0 0.0000000
2015-12-24         0 0.0000000
2015-12-25         1 0.0000000
2015-12-26         0 0.0000000
2015-12-27         0 0.0000000
2015-12-28         0 0.0000000
2015-12-29         0 0.0000000
2015-12-30         0 0.0000000
2015-12-31         0 0.0000000
2016-01-01         0 0.0000000

License

This package is free and open source software, licensed under MIT.