Identify Memory Patterns in Time Series Using Variance Scale Exponent

Methods for calculating the variance scale exponent to identify memory patterns in time series data. Includes tests for white noise, short memory, and long memory. See Fu, H. et al. (2018) .

vse4ts

Introduction

Methods for calculating the variance scale exponent to identify memory patterns in time series data. Includes tests for white noise, short memory, and long memory. See Fu, H. et al. (2018)<doi:10.1016/j.physa.2018.06.092>.

Installation

You can install the development version of vse4ts from GitHub with:

# install.packages("devtools")
devtools::install_github("z-my-cn/vse4ts")

Example

Here is a basic example of how to use the vse function in the vse4ts package:

library(vse4ts)
set.seed(123)
x <- rnorm(1024)
x.vse <- vse(x)
print(x.vse)
#> [1] 0.4987233

This package also provides two hypothesis test functions Wnoise.test and SLmemory.test to test the white noise and short/long memory of a time series, respectively. Here is an example of how to use the Wnoise.test function and SLmemory.test function in the vse4ts package:

library(vse4ts)
# install.packages("pracma")
library(pracma)

data("brown72")
x <- brown72

# Test white noise
Wnoise.test(x)
#> Wnoise Test
#>
#> Wnoise statistic: 135.1091
#> degrees of freedom: 31
#> p-value: 5.884182e-15
#>
#> alternative hypothesis: non-independent stochastic process

# Test long memory
SLmemory.test(x)
#> SLmemory Test
#>
#> SLmemory statistic: 21.20841
#> degrees of freedom: 31
#> p-value: 0.09369624
#>
#> alternative hypothesis: long memory

License

MIT © 2024 vse4ts authors