subgrad condition option added

Author: Jelena Markovic

selectiveInference/R/funs.randomized.R

@@ -258,15 +258,20 @@ conditional_density = function(noise_scale, lasso_soln) {
   return(lasso_soln)
 }
 
-randomized_inference = function(X, y, sigma, lam, noise_scale, ridge_term, level=0.9){
+randomized_inference = function(X, y, sigma, lam, noise_scale, ridge_term, 
+                                condition_subgrad=FALSE, level=0.9){
 
   n = nrow(X)
   p = ncol(X)
   lasso_soln = selectiveInference:::randomizedLASSO(X, y, lam, noise_scale, ridge_term)
   active_set = lasso_soln$active_set
   inactive_set = lasso_soln$inactive_set
   nactive = length(active_set)
-
+  
+  if (condition_subgrad==TRUE){
+    lasso_soln=conditional_density(noise_scale,lasso_soln)
+  } 
+    
   dim = length(lasso_soln$observed_opt_state)
   print(paste("chain dim", dim))
   S = selectiveInference:::sample_opt_variables(lasso_soln, jump_scale=rep(1/sqrt(n), dim), nsample=10000)

tests/randomized/test_instances.R

@@ -25,8 +25,8 @@ gaussian_instance = function(n, p, s, sigma=1, rho=0, signal=6, X=NA,
 }
 
 
-collect_results = function(n,p,s, nsim=10, level=0.9){
-  rho=0.
+collect_results = function(n,p,s, nsim=100, level=0.9){
+  rho=0.3
   lam=1.
   sigma=1
   sample_pvalues = c()
@@ -38,7 +38,7 @@ collect_results = function(n,p,s, nsim=10, level=0.9){
     beta=data$beta
     ridge_term=sd(y)/sqrt(n)
     noise_scale = sd(y)/2
-    result = selectiveInference:::randomized_inference(X,y,sigma,lam,noise_scale,ridge_term, level)
+    result = selectiveInference:::randomized_inference(X,y,sigma,lam,noise_scale,ridge_term, TRUE, level)
     true_beta = beta[result$active_set]
     coverage = rep(0, nrow(result$ci))
     for (i in 1:nrow(result$ci)){