The goal of cusum is to provide functions for constructing and evaluating CUSUM charts and RA-CUSUM charts with focus on false signal probability.
You can install the released version of cusum from CRAN with:
install.packages("cusum")Or the pre-release version from github with:
devtools::install_github("lhubig/cusum")This is a basic example which shows you how to construct CUSUM charts with simulated control limits:
library(cusum)
limit <- cusum_limit_sim(failure_probability = 0.05,
n_patients = 100,
odds_multiplier = 2,
n_simulation = 1000,
alpha = 0.05,
seed = 2018)
print(limit)
#> [1] 2.696187
set.seed(2018)
patient_outcomes <- as.logical(rbinom(n = 100, size = 1, prob = 0.05))
cs <- cusum(failure_probability = 0.05,
patient_outcomes,
limit,
odds_multiplier = 2)
plot(cs)