mergingTools (version 1.0.1)

simulate_and_merge: Simulate the MVG and merge the HEMs

Description

simulate_and_merge takes as input the parameters of the multivariate Gaussian distribution and simulates multiple samples. The number of samples is determined by n_sims. On each simulation, each HEM of the experimental data is sorted from lowest to highest independently. Then, the simulated MGD is used as the model to arrange the experimental data. For instance, if on the MGD the 10th highest value of HEM 1 is paired with the 2nd highest and the 4th highest values of HEMs 2 and 3 respectively, Then, we modify the experimental data to copy this arrangement. After merging and arranging, the 10th highest value of HEM 1 on the experimental data will be paired with the 2nd highest and the 4th highest values of HEMs 2 and 3 respectively. The algorithm does multiple simulations of the MGD and keeps the sample that gives the lowest error on the correlation matrix when the data is merged w.r.t. the correlation matrix of the experimental data.

Usage

simulate_and_merge(
  mvg_params = NULL,
  n_runs = NULL,
  n_sims = NULL,
  cor_matrix = NULL
)

Value

Returns a merged dataframe as stated in the description

Arguments

mvg_params

parameters from the multivariate Gaussian distribution

n_runs

Number of rows for the output merged data

n_sims

Number of simulations for the multivariate Gaussian distribution to find the optimal merge

cor_matrix

Correlation matrix

Examples

n_pmcs <- 6
data_much <- mergingTools::process_raw_experiments(data = data_much_raw_vignette,
                                                   n_pmcs = n_pmcs)
cor_matrix <- mergingTools::correlation_matrix(splitted_data = data_much)
dep_lvl <- 0.85
# Remove the HEMs which are linearly dependant on other HEMs
cor_matrix_independent <- mergingTools::get_independent_matrix(cor_matrix = cor_matrix,
                                                               dep_lvl = dep_lvl)
mvg_params <- mergingTools::generate_mvg_params(splitted_data = data_much,
                                                cor_matrix = cor_matrix_independent)
n_sims <- 10
n_runs <- 1000
merged_data <- mergingTools::simulate_and_merge(mvg_params = mvg_params,
                                                n_runs = n_runs,
                                                n_sims = n_sims,
                                                cor_matrix = cor_matrix_independent)