In ataqi23/RMAT: Random Matrix Analysis Toolkit

knitr::opts_chunk$set(echo = TRUE, warning = FALSE)
library(RMAT)
library(purrr)
source("../R/mat-diag.R")
#source("../R/matrices.R")

Normal Matrices

# Testing parameter accuracy for RM_norm
normal_mean <- function(N = 5, mu = 0, threshold = 0.25){
  P <- RM_norm(N = N, mean = mu)
  mu_sim <- .normal_params(P)[[1]]
  abs(mu - mu_sim) < threshold
}
normal_sd <- function(N, sd = 1, threshold = 0.25){
  P <- RM_norm(N = 5, mean = 0, sd = sd)
  sd_sim <- .normal_params(P)[[2]]
  abs(sd - sd_sim) < threshold
}

# Set parameters
N <- 100
repl <- 100

# Test accuracy
sum(replicate(n = repl, normal_mean(N, mu = 5, threshold = 0.2), simplify = TRUE)) / repl
sum(replicate(repl, normal_sd(N, sd = 1, threshold = 0.2), simplify = TRUE)) / repl

Stochastic Matrices

# Testing parameter accuracy for RM_stoch
stoch_test <- function(N, size, symm = T, sparsity = F){
  ens <- RME_stoch(N = N, symm = symm, sparsity = sparsity, size = size)
  v <- purrr::map_lgl(ens, .isStochastic)  
  return(sum(v) / size) # Get result
}

# Set parameters
N <- 100
size <- 100

# Test accuracy
stoch_test(N = N, size = 1000)
stoch_test(N = N, size = 1000, symm = T)

ataqi23/RMAT documentation built on May 10, 2021, 12:03 a.m.