test_nonlinearity: Randomization test for nonlinearity using S-maps and surrogate data
Description
test_nonlinearity tests for nonlinearity using S-maps by
comparing improvements in forecast skill (delta rho and delta mae) between
linear and nonlinear models with a null distribution from surrogate data.Usage
test_nonlinearity(ts, method = "ebisuzaki", num_surr = 200, T_period = 1,
E = 1, ...)
Arguments
ts
the original time series
method
which algorithm to use to generate surrogate data
num_surr
the number of null surrogates to generate
T_period
the period of seasonality for seasonal surrogates (ignored for other methods)
E
the embedding dimension for s_map
...
optional arguments to s_map
Value
- A data.frame containing the following components:
ll{
delta_rho the value of the delta rho statistic
delta_mae the value of the delat mae statistic
num_surr the size of the null distribution
delta_rho_p_value the p-value for delta rho
delta_mae_p_value the p-value for delta mae
}