22:T66c,# lmmpar [![Travis-CI Build Status](https://travis-ci.org/schloerke/lmmpar.svg?branch=master)](https://travis-ci.org/schloerke/lmmpar) [![CRAN_Status_Badge](http://www.r-pkg.org/badges/version/lmmpar)](https://cran.r-project.org/package=lmmpar) The goal of lmmpar is to ... ## Installation You can install lmmpar from github with: ``` r # install.packages("devtools") devtools::install_github("fulyagokalp/lmmpar") ``` ## Example This is a basic example which shows you how to solve a common problem: ``` r # Set up fake data n <- 10000 # number of subjects m <- 4 # number of repeats N <- n * m # true size of data p <- 50 # number of betas q <- 2 # width of random effects # Initial parameters # beta has a 1 for the first value. all other values are ~N(10, 1) beta <- rbind(1, matrix(rnorm(p, 10), p, 1)) R <- diag(m) D <- matrix(c(16, 0, 0, 0.025), nrow = q) sigma <- 1 # Set up data subject <- rep(1:n, each = m) repeats <- rep(1:m, n) subj_x <- lapply(1:n, function(i) cbind(1, matrix(rnorm(m * p), nrow = m))) X <- do.call(rbind, subj_x) Z <- X[, 1:q] subj_beta <- lapply(1:n, function(i) mnormt::rmnorm(1, rep(0, q), D)) subj_err <- lapply(1:n, function(i) mnormt::rmnorm(1, rep(0, m), sigma * R)) # create a known response subj_y <- lapply( seq_len(n), function(i) { (subj_x[[i]] %*% beta) + (subj_x[[i]][, 1:q] %*% subj_beta[[i]]) + subj_err[[i]] } ) Y <- do.call(rbind, subj_y) # run the algorithm in parallel to recover the known betas ans <- lmmpar( Y, X, Z, subject, beta = beta, R = R, D = D, cores = 4, sigma = sigma, verbose = TRUE ) ```23:T49a,M7.49933 0.25C3.49635 0.25 0.25 3.49593 0.25 7.50024C0.25 10.703 2.32715 13.4206 5.2081 14.3797C5.57084 14.446 5.70302 14.2222 5.70302 14.0299C5.70302 13.8576 5.69679 13.4019 5.69323 12.797C3.67661 13.235 3.25112 11.825 3.25112 11.825C2.92132 10.9874 2.44599 10.7644 2.44599 10.7644C1.78773 10.3149 2.49584 10.3238 2.49584 10.3238C3.22353 10.375 3.60629 11.0711 3.60629 11.0711C4.25298 12

lmmpar0.1.0 package

Functions

Readme

Dependencies

Imports

Versions

News

lmmpar