From e40613944710e6b1a95be8633b599d2d7bac412b Mon Sep 17 00:00:00 2001 From: jajupmochi Date: Thu, 27 Feb 2020 19:21:52 +0100 Subject: [PATCH] update docs --- docs/source/experiments.rst | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/docs/source/experiments.rst b/docs/source/experiments.rst index 151cb54..de137c0 100644 --- a/docs/source/experiments.rst +++ b/docs/source/experiments.rst @@ -1,7 +1,7 @@ Experiments =========== -To exhibit the effectiveness and practicability of `graphkit-learn` library, we tested it on several benchmark datasets. See `(Kersting et al., 2016) __ for details on these datasets. +To exhibit the effectiveness and practicability of `graphkit-learn` library, we tested it on several benchmark datasets. See `(Kersting et al., 2016) `__ for details on these datasets. A two-layer nested cross-validation (CV) is applied to select and evaluate models, where outer CV randomly splits the dataset into 10 folds with 9 as validation set, and inner CV then randomly splits validation set to 10 folds with 9 as training set. The whole procedure is performed 30 times, and the average performance is computed over these trails. Possible parameters of a graph kernel are also tuned during this procedure.