Node2Vec實戰

數據結構

兩個兩個連接的節點

1 2 2 3 4 5

主程序構建

G = nx.read_edgelist('../data/text.txt',create_using=nx.DiGraph(), nodetype=None, data=[('weight', int)]) ## 構建模型 model = Node2Vec(G, walk_length=10, num_walks=80,p=0.25, q=4, workers=1, use_rejection_sampling=0) ## 訓練 model.train(embed_size=4, window_size=5, iter=3) embeddings = model.get_embeddings() print(embeddings)

初始生成節點到節點的概率

def preprocess_transition_probs(self):'''Preprocessing of transition probabilities for guiding the random walks.'''####get_alias_edge這個函數是對每條邊設定為二階randomwalk的概率形式###這個函數的作用是生成每個邊界的概率，同時會有alias_setup這個函數將概率進行轉換，方便后面抽樣G = self.Gis_directed = self.is_directedalias_nodes = {}for node in G.nodes():unnormalized_probs = [G[node][nbr]['weight'] for nbr in sorted(G.neighbors(node))]#讀取每個鄰點權重norm_const = sum(unnormalized_probs)###權重求和，作為公式中正則項常數的那個分母normalized_probs = [float(u_prob)/norm_const for u_prob in unnormalized_probs]###除以分母alias_nodes[node] = alias_setup(normalized_probs)alias_edges = {}triads = {}if is_directed:for edge in G.edges():alias_edges[edge] = self.get_alias_edge(edge[0], edge[1])else:for edge in G.edges():alias_edges[edge] = self.get_alias_edge(edge[0], edge[1])alias_edges[(edge[1], edge[0])] = self.get_alias_edge(edge[1], edge[0])self.alias_nodes = alias_nodesself.alias_edges = alias_edgesreturn

get_alias_edge是得到節點到節點的概率

def get_alias_edge(self, src, dst):####二階ramdom walk#src是隨機游走序列中的上一個節點，dst是當前節點'''Get the alias edge setup lists for a given edge.'''G = self.Gp = self.pq = self.qunnormalized_probs = []for dst_nbr in sorted(G.neighbors(dst)):if dst_nbr == src:unnormalized_probs.append(G[dst][dst_nbr]['weight']/p)elif G.has_edge(dst_nbr, src):unnormalized_probs.append(G[dst][dst_nbr]['weight'])else:unnormalized_probs.append(G[dst][dst_nbr]['weight']/q)norm_const = sum(unnormalized_probs)normalized_probs = [float(u_prob)/norm_const for u_prob in unnormalized_probs]return alias_setup(normalized_probs)

alias_setup ：輸入概率，得到對應的兩組數，方便后面的抽樣調用

def alias_setup(probs):'''alias_setup的作用是根據二階random walk輸出的概率變成每個節點對應兩個數，被后面的alias_draw函數所進行抽樣'''K = len(probs)q = np.zeros(K)J = np.zeros(K, dtype=np.int)smaller = []larger = []for kk, prob in enumerate(probs):q[kk] = K*probif q[kk] < 1.0:smaller.append(kk)else:larger.append(kk)##kk是下標，表示哪些下標小while len(smaller) > 0 and len(larger) > 0:small = smaller.pop()##smaller自己也會減少最右邊的值large = larger.pop()J[small] = largeq[large] = q[large] + q[small] - 1.0if q[large] < 1.0:smaller.append(large)else:larger.append(large)return J, q

alias_draw 抽樣函數

def alias_draw(J, q):'''Draw sample from a non-uniform discrete distribution using alias sampling.'''K = len(J)kk = int(np.floor(np.random.rand()*K))if np.random.rand() < q[kk]:return kkelse:return J[kk]

node2vec_walk就是對于給定的長度，對于開始節點開始模擬這個節點的路徑，涉及的函數都在上面提及

def node2vec_walk(self, walk_length, start_node):'''Simulate a random walk starting from start node.'''G = self.Galias_nodes = self.alias_nodesalias_edges = self.alias_edgeswalk = [start_node]######alias_draw這個函數是等于是根據二階random walk概率選擇下一個點while len(walk) < walk_length:cur = walk[-1]cur_nbrs = sorted(G.neighbors(cur))###G.neighbors(cur)得到cur一級關聯的節點if len(cur_nbrs) > 0:if len(walk) == 1:####cur[0]walk.append(cur_nbrs[alias_draw(alias_nodes[cur][0], alias_nodes[cur][1])])else:prev = walk[-2]next = cur_nbrs[alias_draw(alias_edges[(prev, cur)][0], alias_edges[(prev, cur)][1])]walk.append(next)else:breakreturn walk

總結

以上是生活随笔為你收集整理的Node2Vec实战的全部內容，希望文章能夠幫你解決所遇到的問題。

如果覺得生活随笔網站內容還不錯，歡迎將生活随笔推薦給好友。