Skip to contents

Simulate DCRW Data for Study 1: Different Outlier Configurations

Source: R/simulate_data.R
simulate_data_study1.Rd

Simulates data sets under a first-difference correlated random walk (DCRW) state-space model for Study 1 in the paper. This study evaluates performance under different types of outlier configurations. All arguments have default values matching the simulation setup used in the paper.

Usage

simulate_data_study1(
  sample_sizes = c(100, 200, 500, 1000),
  samples = 100,
  n_oos = 20,
  contamination = 0.1,
  distance = 5,
  sd_cluster = 2,
  mean_cluster = c(20, 20),
  multi_level_distances = c(distance - 2, distance, distance + 2),
  phi_coef = 0.8,
  sigma2_w_lon = 0.1,
  sigma2_w_lat = 0.1,
  sigma2_v_lon = 0.4,
  sigma2_v_lat = 0.4,
  initial_state = c(0, 0, 0, 0),
  seed = NA
)

Arguments

sample_sizes

Vector of sample sizes \(n\) for each simulated dataset. Default is c(100, 200, 500, 1000).

samples

Number of simulated data sets per \(n\) and per outlier configuration. Default is 100.

n_oos

Number of out-of-sample (future) timesteps. Default is 20.

contamination

Proportion of contaminated (outlying) observations. Default is 0.1.

distance

Distance used for fixed-distance outliers. Default is 5.

sd_cluster

Standard deviation of cluster for cluster-based outliers. Default is 2.

mean_cluster

Mean vector of cluster for cluster-based outliers. Default is c(20, 20).

multi_level_distances

Vector of distances for multi-level outliers. Must be of length 3. Default is c(3, 5, 7).

phi_coef

Autocorrelation parameter in the DCRW transition matrix. Ranges between 0 and 1. Default is 0.8.

sigma2_w_lon, sigma2_w_lat

State noise variances (longitude and latitude). Default is 0.1 each.

sigma2_v_lon, sigma2_v_lat

Observation noise variances (longitude and latitude). Default is 0.4 each.

initial_state

Initial state vector of length 4. Default is c(0, 0, 0, 0).

seed

Optional random seed for reproducibility. Default is NA. Use seed = 1302 to reproduce the same data as in the paper.

Value

A tibble containing the simulated data sets. Each row corresponds to a simulated data set and includes fields for sample size, outlier configuration (setting), outliers, clean data, noisy observations, and out-of-sample values.

See also

simulate_data_study2

Examples

data_study1 = simulate_data_study1(samples = 5, seed = 123)