Skip to content

karakastarik/stationarityR

Repository files navigation

stationarityR

Automating Time Series Stationarity Tests

This package automates Kwiatkowski–Phillips–Schmidt–Shin (KPSS), Augmented Dickey–Fuller (ADF) and Phillips–Perron (PP) unit root tests and saves the results as a dataframe.

Getting Started

devtools must be installed to access the stationarityR package. If not installed:

install.packages("devtools", dependencies = TRUE) 

Then you can use following code to install stationarityR package:

library(devtools)
devtools::install_github("karakastarik/stationarityR")
library(stationarityR)

Functions

For 0.1.0 release, only summary_all(model,lag), summary_kpss(model,lag),summary_adf(model,lag) and summary_pp(model,lag) functions are available and these functions returns a dataframe which calculates all possible combinations for unit root tests. In the functions, model is fitted lm(y ~ x) object and lag is integer lag length. For example, if the lag value is 10, results will come for lag lengths of 1:10.

Output

You can see the example dataframe which returned by summary_all(model,lag) function below.

Output The function uses KPSS, ADF and PP unit root tests and decides whether the series is stationary or not according to the lag length, type and significance level (1%, 2.5%, 5% and 10%) as seen above. If the value is pass, it means the series is stationary, and if it is fail, it means the series is not stationary. You can reach this conclusion by examining statistics and critical values(10pct, 5pct...).