Skip to content

Commit 2bdfa07

Browse files
Claudio StamileClaudio Stamile
Claudio Stamile
authored and
Claudio Stamile
committed
Chapter 3 code
1 parent 86d79dd commit 2bdfa07

File tree

4 files changed

+1685
-0
lines changed

4 files changed

+1685
-0
lines changed

‎03_unsupervised_graph_machine_learning/01_Shallow_Embeddings.ipynb

Lines changed: 348 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,348 @@
1+
{
2+
"cells": [
3+
{
4+
"cell_type": "code",
5+
"execution_count": null,
6+
"metadata": {},
7+
"outputs": [],
8+
"source": [
9+
"import matplotlib.pyplot as plt\n",
10+
"\n",
11+
"def draw_graph(G, node_names={}, node_size=500):\n",
12+
" pos_nodes = nx.spring_layout(G)\n",
13+
" nx.draw(G, pos_nodes, with_labels=True, node_size=node_size, edge_color='gray', arrowsize=30)\n",
14+
" \n",
15+
" pos_attrs = {}\n",
16+
" for node, coords in pos_nodes.items():\n",
17+
" pos_attrs[node] = (coords[0], coords[1] + 0.08)\n",
18+
" \n",
19+
" #nx.draw_networkx_labels(G, pos_attrs, font_family='serif', font_size=20)\n",
20+
" \n",
21+
" plt.axis('off')\n",
22+
" axis = plt.gca()\n",
23+
" axis.set_xlim([1.2*x for x in axis.get_xlim()])\n",
24+
" axis.set_ylim([1.2*y for y in axis.get_ylim()])\n",
25+
" plt.show()"
26+
]
27+
},
28+
{
29+
"cell_type": "markdown",
30+
"metadata": {},
31+
"source": [
32+
"## Graph Factorization"
33+
]
34+
},
35+
{
36+
"cell_type": "code",
37+
"execution_count": null,
38+
"metadata": {},
39+
"outputs": [],
40+
"source": [
41+
"import networkx as nx\n",
42+
"\n",
43+
"G = nx.barbell_graph(m1=3, m2=2)\n",
44+
"draw_graph(G)\n"
45+
]
46+
},
47+
{
48+
"cell_type": "code",
49+
"execution_count": null,
50+
"metadata": {},
51+
"outputs": [],
52+
"source": [
53+
"from gem.embedding.gf import GraphFactorization\n",
54+
"\n",
55+
"G = nx.barbell_graph(m1=10, m2=4)\n",
56+
"draw_graph(G)\n",
57+
"\n",
58+
"gf = GraphFactorization(d=2, data_set=None,max_iter=10000, eta=1*10**-4, regu=1.0)\n",
59+
"gf.learn_embedding(G)"
60+
]
61+
},
62+
{
63+
"cell_type": "code",
64+
"execution_count": null,
65+
"metadata": {},
66+
"outputs": [],
67+
"source": [
68+
"import matplotlib.pyplot as plt\n",
69+
"\n",
70+
"fig, ax = plt.subplots(figsize=(10,10))\n",
71+
"\n",
72+
"for x in G.nodes():\n",
73+
" \n",
74+
" v = gf.get_embedding()[x]\n",
75+
" ax.scatter(v[0],v[1], s=1000)\n",
76+
" ax.annotate(str(x), (v[0],v[1]), fontsize=12)"
77+
]
78+
},
79+
{
80+
"cell_type": "markdown",
81+
"metadata": {},
82+
"source": [
83+
"## GraphRep"
84+
]
85+
},
86+
{
87+
"cell_type": "code",
88+
"execution_count": null,
89+
"metadata": {},
90+
"outputs": [],
91+
"source": [
92+
"import networkx as nx\n",
93+
"from karateclub.node_embedding.neighbourhood.grarep import GraRep\n",
94+
"\n",
95+
"G = nx.barbell_graph(m1=10, m2=4)\n",
96+
"draw_graph(G)\n",
97+
"\n",
98+
"gr = GraRep(dimensions=2,order=3)\n",
99+
"gr.fit(G)"
100+
]
101+
},
102+
{
103+
"cell_type": "code",
104+
"execution_count": null,
105+
"metadata": {},
106+
"outputs": [],
107+
"source": [
108+
"import matplotlib.pyplot as plt\n",
109+
"\n",
110+
"fig, ax = plt.subplots(figsize=(10,10))\n",
111+
"\n",
112+
"ida = 4\n",
113+
"idb = 5\n",
114+
"for x in G.nodes():\n",
115+
" \n",
116+
" v = gr.get_embedding()[x]\n",
117+
" ax.scatter(v[ida],v[idb], s=1000)\n",
118+
" ax.annotate(str(x), (v[ida],v[idb]), fontsize=12)"
119+
]
120+
},
121+
{
122+
"cell_type": "markdown",
123+
"metadata": {},
124+
"source": [
125+
"## HOPE"
126+
]
127+
},
128+
{
129+
"cell_type": "code",
130+
"execution_count": null,
131+
"metadata": {},
132+
"outputs": [],
133+
"source": [
134+
"import networkx as nx\n",
135+
"from gem.embedding.hope import HOPE\n",
136+
"\n",
137+
"G = nx.barbell_graph(m1=10, m2=4)\n",
138+
"draw_graph(G)\n",
139+
"\n",
140+
"hp = HOPE(d=4, beta=0.01)\n",
141+
"hp.learn_embedding(G)"
142+
]
143+
},
144+
{
145+
"cell_type": "code",
146+
"execution_count": null,
147+
"metadata": {},
148+
"outputs": [],
149+
"source": [
150+
"import matplotlib.pyplot as plt\n",
151+
"\n",
152+
"fig, ax = plt.subplots(figsize=(10,10))\n",
153+
"\n",
154+
"for x in G.nodes():\n",
155+
" \n",
156+
" v = hp.get_embedding()[x,2:]\n",
157+
" ax.scatter(v[0],v[1], s=1000)\n",
158+
" ax.annotate(str(x), (v[0],v[1]), fontsize=20)"
159+
]
160+
},
161+
{
162+
"cell_type": "markdown",
163+
"metadata": {},
164+
"source": [
165+
"## DeepWalk"
166+
]
167+
},
168+
{
169+
"cell_type": "code",
170+
"execution_count": null,
171+
"metadata": {},
172+
"outputs": [],
173+
"source": [
174+
"import networkx as nx\n",
175+
"from karateclub.node_embedding.neighbourhood.deepwalk import DeepWalk\n",
176+
"\n",
177+
"G = nx.barbell_graph(m1=10, m2=4)\n",
178+
"draw_graph(G)\n",
179+
"\n",
180+
"dw = DeepWalk(dimensions=2)\n",
181+
"dw.fit(G)"
182+
]
183+
},
184+
{
185+
"cell_type": "code",
186+
"execution_count": null,
187+
"metadata": {},
188+
"outputs": [],
189+
"source": [
190+
"import matplotlib.pyplot as plt\n",
191+
"\n",
192+
"fig, ax = plt.subplots(figsize=(10,10))\n",
193+
"\n",
194+
"for x in G.nodes():\n",
195+
" \n",
196+
" v = dw.get_embedding()[x]\n",
197+
" ax.scatter(v[0],v[1], s=1000)\n",
198+
" ax.annotate(str(x), (v[0],v[1]), fontsize=12)"
199+
]
200+
},
201+
{
202+
"cell_type": "markdown",
203+
"metadata": {},
204+
"source": [
205+
"## Node2Vec"
206+
]
207+
},
208+
{
209+
"cell_type": "code",
210+
"execution_count": null,
211+
"metadata": {},
212+
"outputs": [],
213+
"source": [
214+
"import networkx as nx\n",
215+
"from node2vec import Node2Vec\n",
216+
"\n",
217+
"G = nx.barbell_graph(m1=10, m2=4)\n",
218+
"draw_graph(G)\n",
219+
"\n",
220+
"node2vec = Node2Vec(G, dimensions=2)\n",
221+
"model = node2vec.fit(window=10)\n",
222+
"embeddings = model.wv"
223+
]
224+
},
225+
{
226+
"cell_type": "code",
227+
"execution_count": null,
228+
"metadata": {},
229+
"outputs": [],
230+
"source": [
231+
"import matplotlib.pyplot as plt\n",
232+
"\n",
233+
"fig, ax = plt.subplots(figsize=(10,10))\n",
234+
"\n",
235+
"for x in G.nodes():\n",
236+
" \n",
237+
" v = model.wv[str(x)]\n",
238+
" ax.scatter(v[0],v[1], s=1000)\n",
239+
" ax.annotate(str(x), (v[0],v[1]), fontsize=16)\n",
240+
"\n",
241+
"plt.show()"
242+
]
243+
},
244+
{
245+
"cell_type": "markdown",
246+
"metadata": {},
247+
"source": [
248+
"## Edge2Vec"
249+
]
250+
},
251+
{
252+
"cell_type": "code",
253+
"execution_count": null,
254+
"metadata": {},
255+
"outputs": [],
256+
"source": [
257+
"from node2vec.edges import HadamardEmbedder\n",
258+
"edges_embs = HadamardEmbedder(keyed_vectors=model.wv)"
259+
]
260+
},
261+
{
262+
"cell_type": "code",
263+
"execution_count": null,
264+
"metadata": {},
265+
"outputs": [],
266+
"source": [
267+
"import matplotlib.pyplot as plt\n",
268+
"\n",
269+
"fig, ax = plt.subplots(figsize=(10,10))\n",
270+
"\n",
271+
"for x in G.edges():\n",
272+
" \n",
273+
" v = edges_embs[(str(x[0]), str(x[1]))]\n",
274+
" ax.scatter(v[0],v[1], s=1000)\n",
275+
" ax.annotate(str(x), (v[0],v[1]), fontsize=16)\n",
276+
"\n",
277+
"plt.show()"
278+
]
279+
},
280+
{
281+
"cell_type": "markdown",
282+
"metadata": {},
283+
"source": [
284+
"## Graph2Vec"
285+
]
286+
},
287+
{
288+
"cell_type": "code",
289+
"execution_count": null,
290+
"metadata": {},
291+
"outputs": [],
292+
"source": [
293+
"import random\n",
294+
"import matplotlib.pyplot as plt\n",
295+
"from karateclub import Graph2Vec\n",
296+
"\n",
297+
"n_graphs = 20\n",
298+
"\n",
299+
"def generate_radom():\n",
300+
" n = random.randint(6, 20)\n",
301+
" k = random.randint(5, n)\n",
302+
" p = random.uniform(0, 1)\n",
303+
" return nx.watts_strogatz_graph(n,k,p), [n,k,p]\n",
304+
"\n",
305+
"Gs = [generate_radom() for x in range(n_graphs)]\n",
306+
"\n",
307+
"model = Graph2Vec(dimensions=2, wl_iterations=10)\n",
308+
"model.fit([x[0] for x in Gs])\n",
309+
"embeddings = model.get_embedding()\n",
310+
"\n",
311+
"fig, ax = plt.subplots(figsize=(10,10))\n",
312+
"\n",
313+
"for i,vec in enumerate(embeddings):\n",
314+
" \n",
315+
" ax.scatter(vec[0],vec[1], s=1000)\n",
316+
" ax.annotate(str(i), (vec[0],vec[1]), fontsize=40)"
317+
]
318+
},
319+
{
320+
"cell_type": "code",
321+
"execution_count": null,
322+
"metadata": {},
323+
"outputs": [],
324+
"source": []
325+
}
326+
],
327+
"metadata": {
328+
"kernelspec": {
329+
"display_name": "Python 3",
330+
"language": "python",
331+
"name": "python3"
332+
},
333+
"language_info": {
334+
"codemirror_mode": {
335+
"name": "ipython",
336+
"version": 3
337+
},
338+
"file_extension": ".py",
339+
"mimetype": "text/x-python",
340+
"name": "python",
341+
"nbconvert_exporter": "python",
342+
"pygments_lexer": "ipython3",
343+
"version": "3.6.8"
344+
}
345+
},
346+
"nbformat": 4,
347+
"nbformat_minor": 4
348+
}

0 commit comments

Comments
 (0)