From 2d41a7c807a9f515649cce77c3268aab7f997a60 Mon Sep 17 00:00:00 2001
From: Wapiti <tzrzhuoran@gmail.com>
Date: Mon, 21 Feb 2022 14:46:19 +0000
Subject: [PATCH] Update 01_Social_network_analysis.ipynb

Should based on the whole Graph to generate the model:

Have confirmed with the following experiment:

Node2Vec for graph_train:
  - Precision: 0.9033816425120773
  - Recall: 0.9091249065071054
  - F1-Score: 0.9062441752096925

Node2Vec for G:
  - Precision: 0.9235116804822909
  - Recall: 0.9166043380703066
  - F1-Score: 0.920045045045045
---
 Chapter06/01_Social_network_analysis.ipynb | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/Chapter06/01_Social_network_analysis.ipynb b/Chapter06/01_Social_network_analysis.ipynb
index 535a774..2455a71 100644
--- a/Chapter06/01_Social_network_analysis.ipynb
+++ b/Chapter06/01_Social_network_analysis.ipynb
@@ -1207,7 +1207,7 @@
     "from node2vec.edges import HadamardEmbedder \n",
     "from stellargraph.data import EdgeSplitter \n",
     "\n",
-    "node2vec = Node2Vec(graph_train) \n",
+    "node2vec = Node2Vec(G) \n",
     "model = node2vec.fit() \n",
     "edges_embs = HadamardEmbedder(keyed_vectors=model.wv) \n",
     "train_embeddings = [edges_embs[str(x[0]),str(x[1])] for x in samples_train]\n",