使用DeepWalk從圖中提取特征

import networkx as nximport pandas as pdimport numpy as npimport randomfrom tqdm import tqdmfrom sklearn.decomposition import PCA
import matplotlib.pyplot as plt%matplotlib inline

df = pd.read_csv("space_data.tsv", sep = "\t")df.head()

G = nx.from_pandas_edgelist(df, "source", "target", edge_attr=True, create_using=nx.Graph())

len(G)

def get_randomwalk(node, path_length):
    random_walk = [node]
    for i in range(path_length-1):        temp = list(G.neighbors(node))        temp = list(set(temp) - set(random_walk))            if len(temp) == 0:            break
        random_node = random.choice(temp)        random_walk.append(random_node)        node = random_node
    return random_walk

get_randomwalk('space exploration', 10)

# 從圖獲取所有節(jié)點的列表all_nodes = list(G.nodes())
random_walks = []for n in tqdm(all_nodes):    for i in range(5):        random_walks.append(get_randomwalk(n,10))
# 序列個數(shù)len(random_walks)

# importing required libraries
import pandas as pdimport networkx as nximport numpy as npimport randomfrom tqdm import tqdmfrom sklearn.decomposition import PCAimport pprintfrom gensim.models import Word2Vecimport warningswarnings.filterwarnings('ignore')
# read the datasetdf = pd.read_csv("space_data.tsv", sep = "\t")print(df.head())
G = nx.from_pandas_edgelist(df, "source", "target", edge_attr=True, create_using=nx.Graph())
G = nx.from_pandas_edgelist(df, "source", "target", edge_attr=True, create_using=nx.Graph())
print('The number of nodes in pur graph: ',len(G))
def get_randomwalk(node, path_length):        random_walk = [node]        for i in range(path_length-1):        temp = list(G.neighbors(node))        temp = list(set(temp) - set(random_walk))            if len(temp) == 0:            break
        random_node = random.choice(temp)        random_walk.append(random_node)        node = random_node            return random_walk
print('\n\nRandom sequence of nodes generated from Random Walk\n\n')while True:  first_node = input("Enter name of first node (for example 'space exploration') : ")  if len(first_node) > 0:    breakpprint.pprint(get_randomwalk(first_node, 10))


# 從圖中獲取所有節(jié)點的列表all_nodes = list(G.nodes())
random_walks = []for n in tqdm(all_nodes):    for i in range(5):        random_walks.append(get_randomwalk(n,10))        # 序列長度len(random_walks)

# 訓(xùn)練skip-gram (word2vec)模型model = Word2Vec(window = 4, sg = 1, hs = 0,                 negative = 10, # 負(fù)采樣                 alpha=0.03, min_alpha=0.0007,                 seed = 14)
model.build_vocab(random_walks, progress_per=2)
model.train(random_walks, total_examples = model.corpus_count, epochs=20, report_delay=1)print('\n\n Get similar nodes\n\n')while True:  any_node = input("Enter name of any node (for example 'space toursim') : ")  if len(any_node) > 0:    breakpprint.pprint(model.similar_by_word(any_node))