From 21f9c9ef59a697fe93c3eb67382fae9dfe9fc6a6 Mon Sep 17 00:00:00 2001
From: ihar <Ihar.Shulhan@gmail.com>
Date: Tue, 17 Nov 2020 11:01:53 +0100
Subject: [PATCH] Fix an issue with nan mean_shadow

---
 boostaroota/boostaroota.py | 1 +
 1 file changed, 1 insertion(+)

diff --git a/boostaroota/boostaroota.py b/boostaroota/boostaroota.py
index c1712d1..f60b979 100644
--- a/boostaroota/boostaroota.py
+++ b/boostaroota/boostaroota.py
@@ -134,6 +134,7 @@ def _reduce_vars_xgb(x, y, metric, this_round, cutoff, n_iterations, delta, sile
 
     # Get mean value from the shadows
     mean_shadow = shadow_vars['Mean'].mean() / cutoff
+    mean_shadow = mean_shadow if not np.isnan(mean_shadow) else 0
     real_vars = real_vars[(real_vars.Mean > mean_shadow)]
 
     #Check for the stopping criteria