Merge branch 'release-1.0.0'

.Rbuildignore

@@ -5,3 +5,5 @@ CITATION.cff
 aux/
 ^\.github$
 ^.*\.out
+^cran-comments\.md$
+^CRAN-SUBMISSION$

.github/workflows/R-CMD-check.yaml

@@ -2,7 +2,9 @@
 # Need help debugging build failures? Start at https://github.com/r-lib/actions#where-to-find-help
 on:
   push:
-    branches: [main]
+    branches:
+      - main
+      - 'release-*'
   pull_request:
     branches: [main]
 

CRAN-SUBMISSION

@@ -0,0 +1,3 @@
+Version: 1.0.0
+Date: 2025-01-08 12:10:09 UTC
+SHA: af2656631394df0b8771edb9d2ebcacdad682896

DESCRIPTION

@@ -1,6 +1,6 @@
 Package: MADMMplasso
-Title: Multi Variate Multi Response 'ADMM' with Interaction Effects
-Version: 0.0.0.9024
+Title: Multi Variate Multi Response ADMM with Interaction Effects
+Version: 1.0.0
 Authors@R:
     c(
         person(
@@ -25,9 +25,13 @@ Authors@R:
             role = "aut"
         )
     )
-Description: This system allows one to model a multi response problem
-    that has correlation among the responses and also has main and interaction
-    effects among the covariates. The implementation is based on the methodology
+Description: This system allows one to model a multi-variate, multi-response
+    problem with interaction effects. It combines the usual squared error loss
+    for the multi-response problem with some penalty terms to encourage
+    responses that correlate to form groups and also allow for modeling main and
+    interaction effects that exit within the covariates.
+    The optimization method employed is the Alternating Direction Method of
+    Multipliers (ADMM). The implementation is based on the methodology
     presented on Quachie Asenso, T., & Zucknick, M. (2023)
     <doi:10.48550/arXiv.2303.11155>.
 Imports:
@@ -46,10 +50,11 @@ Imports:
 LinkingTo: Rcpp, RcppArmadillo
 Suggests:
     testthat (>= 3.0.0),
-    pracma,
     lintr
 License: GPL-3
 Encoding: UTF-8
 Roxygen: list(markdown = TRUE)
 RoxygenNote: 7.3.2
 Config/testthat/edition: 3
+Date: 2025-01-08
+Language: en-GB

NEWS.md

@@ -0,0 +1,3 @@
+# MADMMplasso 1.0.0
+
+* Initial CRAN submission.

R/MADMMplasso.R

@@ -12,7 +12,7 @@
 #' @param rho the Lagrange variable for the ADMM. This value is updated during the ADMM call based on a certain condition.
 #' @param e.abs absolute error for the ADMM
 #' @param e.rel relative error for the ADMM
-#' @param gg penalty term for the tree structure. This is a 2x2 matrix values in the first row representing the maximum to the minimum values for lambda_1 and the second row representing the maximum to the minimum values for lambda_2. In the current setting, we set both maximum and the minimum to be same because cross validation is not carried across the lambda_1 and lambda_2. However, setting different values will work during the model fit.
+#' @param gg penalty term for the tree structure. This is a 2×2 matrix values in the first row representing the maximum to the minimum values for lambda_1 and the second row representing the maximum to the minimum values for lambda_2. In the current setting, we set both maximum and the minimum to be same because cross validation is not carried across the lambda_1 and lambda_2. However, setting different values will work during the model fit.
 #' @param my_lambda user specified lambda_3 values
 #' @param lambda_min the smallest value for lambda_3 , as a fraction of max(lambda_3), the (data derived (lammax)) entry value (i.e. the smallest value for which all coefficients are zero)
 #' @param max_it 	maximum number of iterations in loop for one lambda during the ADMM optimization

R/RcppExports.R

@@ -5,7 +5,7 @@
 #' @param X  n by p matrix of predictors
 #' @param Z n by nz matrix of modifying variables. The elements of z
 #' may represent quantitative or categorical variables, or a mixture of the two.
-#' Categorical varables should be coded by 0-1 dummy variables: for a k-level
+#' Categorical variables should be coded by 0-1 dummy variables: for a k-level
 #' variable, one can use either k or k-1  dummy variables.
 #' @param beta0 a vector of length ncol(y) of estimated beta_0 coefficients
 #' @param theta0 matrix of the initial theta_0 coefficients  ncol(Z) by ncol(y)
@@ -18,10 +18,10 @@
 #' @param XtY a matrix formed by multiplying the transpose of X by y.
 #' @param y  N by D matrix  of responses. The X and Z variables are centered in the function. We recommend that X and Z also be standardized before the call
 #' @param N nrow(X)
-#' @param e_abs absolute error for the admm. This is included int the call of MADMMplasso.
-#' @param e_rel relative error for the admm. This is included int the call of MADMMplasso.
+#' @param e_abs absolute error for the ADMM. This is included int the call of MADMMplasso.
+#' @param e_rel relative error for the ADMM. This is included int the call of MADMMplasso.
 #' @param alpha mixing parameter, usually obtained from the MADMMplasso call. When the goal is to include more interactions, alpha should be very small and vice versa.
-#' @param lambda a vector  lambda_3 values for the admm call with length ncol(y). This is usually calculated in the MADMMplasso call.   In our current setting, we use the same the lambda_3 value for all responses.
+#' @param lambda a vector  lambda_3 values for the ADMM call with length ncol(y). This is usually calculated in the MADMMplasso call.   In our current setting, we use the same the lambda_3 value for all responses.
 #' @param alph an overrelaxation parameter in \[1, 1.8\], usually obtained from the MADMMplasso call.
 #' @param svd_w_tu the transpose of the U matrix from the SVD of W_hat
 #' @param svd_w_tv the transpose of the V matrix from the SVD of W_hat
@@ -31,7 +31,7 @@
 #' The easy way to obtain this is by using the function (tree_parms) which gives a default clustering.
 #' However, user decide on a specific structure and then input a tree that follows such structure.
 #' @param my_print Should information form each ADMM iteration be printed along the way? Default TRUE. This prints  the dual and primal residuals
-#' @param gg penalty terms for the tree structure for lambda_1 and  lambda_2 for the admm call.
+#' @param gg penalty terms for the tree structure for lambda_1 and  lambda_2 for the ADMM call.
 #' @return  predicted values for the ADMM part
 #' @description This function fits a multi-response pliable lasso model over a path of regularization values.
 #' @export

R/admm_MADMMplasso.R

@@ -13,7 +13,7 @@
 #' @param N nrow(X)
 #' @param svd.w singular value decomposition of W
 #' @param invmat A list of length ncol(y), each containing the C_d part of equation 32 in the paper
-#' @param gg penalty terms for the tree structure for lambda_1 and  lambda_2 for the admm call.
+#' @param gg penalty terms for the tree structure for lambda_1 and  lambda_2 for the ADMM call.
 #' @return  predicted values for the ADMM part
 
 #' beta0:  estimated beta_0 coefficients  having a size of 1 by ncol(y)

R/cv_MADMMplasso.R

@@ -25,7 +25,7 @@ cv_MADMMplasso <- function(fit, nfolds, X, Z, y, alpha = 0.5, lambda = fit$Lambd
   nfolds <- length(table(foldid))
 
   for (ii in 1:nfolds) {
-    cat("fold,", ii)
+    if (my_print) cat("fold,", ii)
     oo <- foldid == ii
 
     ggg[[ii]] <- MADMMplasso(X = X[!oo, , drop = FALSE], Z = Z[!oo, , drop = FALSE], y = y[!oo, , drop = FALSE], alpha = alpha, my_lambda = lambda, lambda_min = 0.01, max_it = max_it, e.abs = e.abs, e.rel = e.rel, nlambda = length(lambda[, 1]), rho = rho, tree = TT, my_print = my_print, alph = alph, pal = pal, gg = gg, tol = tol, cl = cl, legacy)

R/predict.MADMMplasso.R

@@ -10,8 +10,7 @@
 #' @param y N by D matrix  of responses.
 #' @param lambda  values of lambda at which predictions are desired. If NULL (default), the path of lambda values from the fitted model. are used. If lambda is not NULL, the predictions are made at the closest values to lambda in the lambda path from the fitted model
 #' @param ... additional arguments to the generic \code{predict()} method
-
-#'  @return  predicted values
+#' @return  predicted values
 #' @export
 predict.MADMMplasso <- function(object, X, Z, y, lambda = NULL, ...) {
   lambda.arg <- lambda

R/tree_parms.R

@@ -3,7 +3,7 @@
 #' @param y  N by D matrix of response variables
 #' @param h is the tree cut off
 #' @return  A trained  tree with the following components:
-#' Tree: the tree matrix stored in 1's and 0's
+#' Tree: the tree matrix stored in 1s and 0s
 #'  Tw: tree weights associated with the tree matrix. Each weight corresponds to a row in the tree matrix.
 #'  h_clust: Summary of the hclust call
 #'  y.colnames: names of the response

README.md

@@ -2,8 +2,8 @@
 
 Multi variate multi-response 'ADMM' with interaction effects combines the usual squared error loss for the mult-response problem with some penalty terms  to encourage responses that correlate to form groups and also allow for modeling main and interaction effects that exit within the covariates.
 
-The method can be powperful in situations where one assumes that;
-1.	certain factors influence the main covariate seperatly and aims to include these fatcors as modifying varibles to the main covariate.
+The method can be powerful in situations where one assumes that;
+1.	certain factors influence the main covariate separately and aims to include these factors as modifying variables to the main covariate.
 2.	There exists some form of grouping within the responses and want to include this information. We assume that the responses form overlapping groups that follows a certain hierarchy.
 A typical example is when one wants to model drug response for multiple drugs and assumes that some of the drugs share certain properties in common, for example drug target and chemical compounds and aims to include this information to improve prediction and also aim to predict which drug could be suitable for which patient (given a particular disease). The various diseases under study could be the modifying variable.
 
@@ -65,7 +65,7 @@ TT <- tree_parms(y)
 plot(TT$h_clust)
 ```
 
-![githubb](https://github.com/ocbe-uio/MADMMplasso/assets/85598983/1a843b46-7154-405c-8db6-cec5b7a0982d)
+![github](https://github.com/ocbe-uio/MADMMplasso/assets/85598983/1a843b46-7154-405c-8db6-cec5b7a0982d)
 
 ```r
 gg1 <- matrix(0,2,2)

cran-comments.md

@@ -0,0 +1,13 @@
+Resubmission of 1.0.0, addressing all points from reviewer. Namely:
+
+- Single quotes around ADMM have been removed
+- ADMM acronym has been explained in the package description
+- \value tag has been added to predict.MADMMplasso()
+- Wrapped examples have been unwrapped
+- cat() calls have been conditioned to a verbose argument
+
+## R CMD check results
+
+0 errors | 0 warnings | 1 note
+
+* This is a new release.

inst/examples/MADMMplasso_example.R

@@ -82,7 +82,7 @@ tol <- 1E-3
 fit <- MADMMplasso(
   X, Z, y,
   alpha = alpha, my_lambda = matrix(rep(0.2, ncol(y)), 1),
-  lambda_min = 0.001, max_it = 5000, e.abs = e.abs, e.rel = e.rel, maxgrid = nlambda,
-  nlambda = nlambda, rho = 5, tree = TT, my_print = FALSE, alph = TRUE,
-  pal = TRUE, gg = gg1, tol = tol, cl = 6
+  lambda_min = 0.001, max_it = 1000, e.abs = e.abs, e.rel = e.rel,
+  maxgrid = nlambda, nlambda = nlambda, rho = 5, tree = TT, my_print = FALSE,
+  alph = TRUE, gg = gg1, tol = tol, cl = 2L
 )

inst/examples/cv_MADMMplasso_example.R

@@ -1,97 +1,91 @@
- # nolint start: indentation_linter
-\dontrun{
-  # Train the model
-  # generate some data
-  set.seed(1235)
-  N <- 100
-  p <- 50
-  nz <- 4
-  K <- nz
-  X <- matrix(rnorm(n = N * p), nrow = N, ncol = p)
-  mx <- colMeans(X)
-  sx <- sqrt(apply(X, 2, var))
-  X <- scale(X, mx, sx)
-  X <- matrix(as.numeric(X), N, p)
-  Z <- matrix(rnorm(N * nz), N, nz)
-  mz <- colMeans(Z)
-  sz <- sqrt(apply(Z, 2, var))
-  Z <- scale(Z, mz, sz)
-  beta_1 <- rep(x = 0, times = p)
-  beta_2 <- rep(x = 0, times = p)
-  beta_3 <- rep(x = 0, times = p)
-  beta_4 <- rep(x = 0, times = p)
-  beta_5 <- rep(x = 0, times = p)
-  beta_6 <- rep(x = 0, times = p)
+# Train the model
+# generate some data
+set.seed(1235)
+N <- 100
+p <- 50
+nz <- 4
+K <- nz
+X <- matrix(rnorm(n = N * p), nrow = N, ncol = p)
+mx <- colMeans(X)
+sx <- sqrt(apply(X, 2, var))
+X <- scale(X, mx, sx)
+X <- matrix(as.numeric(X), N, p)
+Z <- matrix(rnorm(N * nz), N, nz)
+mz <- colMeans(Z)
+sz <- sqrt(apply(Z, 2, var))
+Z <- scale(Z, mz, sz)
+beta_1 <- rep(x = 0, times = p)
+beta_2 <- rep(x = 0, times = p)
+beta_3 <- rep(x = 0, times = p)
+beta_4 <- rep(x = 0, times = p)
+beta_5 <- rep(x = 0, times = p)
+beta_6 <- rep(x = 0, times = p)
 
-  beta_1[1:5] <- c(2, 2, 2, 2, 2)
-  beta_2[1:5] <- c(2, 2, 2, 2, 2)
-  beta_3[6:10] <- c(2, 2, 2, -2, -2)
-  beta_4[6:10] <- c(2, 2, 2, -2, -2)
-  beta_5[11:15] <- c(-2, -2, -2, -2, -2)
-  beta_6[11:15] <- c(-2, -2, -2, -2, -2)
+beta_1[1:5] <- c(2, 2, 2, 2, 2)
+beta_2[1:5] <- c(2, 2, 2, 2, 2)
+beta_3[6:10] <- c(2, 2, 2, -2, -2)
+beta_4[6:10] <- c(2, 2, 2, -2, -2)
+beta_5[11:15] <- c(-2, -2, -2, -2, -2)
+beta_6[11:15] <- c(-2, -2, -2, -2, -2)
 
-  Beta <- cbind(beta_1, beta_2, beta_3, beta_4, beta_5, beta_6)
-  colnames(Beta) <- c(1:6)
+Beta <- cbind(beta_1, beta_2, beta_3, beta_4, beta_5, beta_6)
+colnames(Beta) <- c(1:6)
 
-  theta <- array(0, c(p, K, 6))
-  theta[1, 1, 1] <- 2
-  theta[3, 2, 1] <- 2
-  theta[4, 3, 1] <- -2
-  theta[5, 4, 1] <- -2
-  theta[1, 1, 2] <- 2
-  theta[3, 2, 2] <- 2
-  theta[4, 3, 2] <- -2
-  theta[5, 4, 2] <- -2
-  theta[6, 1, 3] <- 2
-  theta[8, 2, 3] <- 2
-  theta[9, 3, 3] <- -2
-  theta[10, 4, 3] <- -2
-  theta[6, 1, 4] <- 2
-  theta[8, 2, 4] <- 2
-  theta[9, 3, 4] <- -2
-  theta[10, 4, 4] <- -2
-  theta[11, 1, 5] <- 2
-  theta[13, 2, 5] <- 2
-  theta[14, 3, 5] <- -2
-  theta[15, 4, 5] <- -2
-  theta[11, 1, 6] <- 2
-  theta[13, 2, 6] <- 2
-  theta[14, 3, 6] <- -2
-  theta[15, 4, 6] <- -2
+theta <- array(0, c(p, K, 6))
+theta[1, 1, 1] <- 2
+theta[3, 2, 1] <- 2
+theta[4, 3, 1] <- -2
+theta[5, 4, 1] <- -2
+theta[1, 1, 2] <- 2
+theta[3, 2, 2] <- 2
+theta[4, 3, 2] <- -2
+theta[5, 4, 2] <- -2
+theta[6, 1, 3] <- 2
+theta[8, 2, 3] <- 2
+theta[9, 3, 3] <- -2
+theta[10, 4, 3] <- -2
+theta[6, 1, 4] <- 2
+theta[8, 2, 4] <- 2
+theta[9, 3, 4] <- -2
+theta[10, 4, 4] <- -2
+theta[11, 1, 5] <- 2
+theta[13, 2, 5] <- 2
+theta[14, 3, 5] <- -2
+theta[15, 4, 5] <- -2
+theta[11, 1, 6] <- 2
+theta[13, 2, 6] <- 2
+theta[14, 3, 6] <- -2
+theta[15, 4, 6] <- -2
 
-  pliable = matrix(0,N,6)
-  for (e in 1:6) {
-    pliable[,e]<-	compute_pliable(X, Z, theta[,,e])
-  }
-
-  esd<-diag(6)
-  e<-MASS::mvrnorm(N,mu=rep(0,6),Sigma=esd)
-  y_train<-X%*%Beta+pliable+e
-  y=y_train
+pliable <- matrix(0, N, 6)
+for (e in 1:6) {
+  pliable[, e] <- compute_pliable(X, Z, theta[, , e])
+}
 
-  colnames(y)<- c( paste("y",1:(ncol(y)),sep = "") )
-  TT=tree_parms(y)
-  plot(TT$h_clust)
-  gg1=matrix(0,2,2)
-  gg1[1,]<-c(0.02,0.02)
-  gg1[2,]<-c(0.02,0.02)
-  nlambda = 50
-  e.abs=1E-4
-  e.rel=1E-2
-  alpha=.2
-  tol=1E-3
-  fit <- MADMMplasso(
-    X, Z, y, alpha=alpha, my_lambda=NULL, lambda_min=0.001, max_it=5000,
-    e.abs=e.abs, e.rel=e.rel, maxgrid=50, nlambda=nlambda, rho=5,tree=TT,
-    my_print=FALSE, alph=1, parallel=FALSE, pal=TRUE, gg=gg1, tol=tol, cl=6
-  )
-  gg1=fit$gg
+esd <- diag(6)
+e <- MASS::mvrnorm(N, mu = rep(0, 6), Sigma = esd)
+y_train <- X %*% Beta + pliable + e
+y <- y_train
 
-  cv_admp <- cv_MADMMplasso(
-    fit, nfolds=5, X, Z, y, alpha=alpha, lambda=fit$Lambdas, max_it=5000,
-    e.abs=e.abs, e.rel=e.rel, nlambda, rho=5, my_print=FALSE, alph=1,
-    foldid=NULL, parallel=FALSE, pal=TRUE, gg=gg1, TT=TT, tol=tol
-  )
-  plot(cv_admp)
-}
-# nolint end: indentation_linter
+colnames(y) <- c(paste("y", 1:(ncol(y)), sep = ""))
+TT <- tree_parms(y)
+plot(TT$h_clust)
+gg1 <- matrix(0, 2, 2)
+gg1[1, ] <- c(0.02, 0.02)
+gg1[2, ] <- c(0.02, 0.02)
+nlambda <- 3
+e.abs <- 1E-3
+e.rel <- 1E-1
+alpha <- .2
+tol <- 1E-2
+fit <- MADMMplasso(
+  X, Z, y, alpha = alpha, my_lambda = NULL, lambda_min = 0.001, max_it = 100,
+  e.abs = e.abs, e.rel = e.rel, maxgrid = nlambda, nlambda = nlambda, rho = 5,
+  tree = TT, my_print = FALSE, alph = 1, gg = gg1, tol = tol, cl = 2L
+)
+cv_admp <- cv_MADMMplasso(
+  fit, nfolds = 5, X, Z, y, alpha = alpha, lambda = fit$Lambdas, max_it = 100,
+  e.abs = e.abs, e.rel = e.rel, nlambda, rho = 5, my_print = FALSE, alph = 1,
+  foldid = NULL, gg = fit$gg, TT = TT, tol = tol
+)
+plot(cv_admp)