如何用python编写链接预测精度评估代码？

pip install networkx

import networkx as nx
from sklearn.metrics import roc_auc_score, precision_recall_fscore_support
import numpy as np

# 创建一个示例图
G = nx.karate_club_graph()

# 分割图为训练集和测试集
def split_graph(G, train_frac=0.8):
    edges = list(G.edges())
    np.random.shuffle(edges)
    split_idx = int(train_frac * len(edges))
    train_edges = edges[:split_idx]
    test_edges = edges[split_idx:]
    return train_edges, test_edges

train_edges, test_edges = split_graph(G)

# 训练集图和测试集图
G_train = G.copy()
G_train.remove_edges_from(test_edges)
G_test = nx.Graph()
G_test.add_edges_from(test_edges)

# Jaccard系数作为链接预测方法
def jaccard_coefficient(u, v, G):
    if u in G and v in G[u]:
        return 1.0
    union_size = len(set(G[u]) | set(G[v]))
    if union_size == 0:
        return 0.0
    return len(set(G[u]) & set(G[v])) / union_size

# 预测所有非边的分数
y_true = []
y_score = []
for u in G.nodes():
    for v in G.nodes():
        if u >= v or (u, v) in train_edges or (v, u) in train_edges:
            continue
        score = jaccard_coefficient(u, v, G_train)
        y_true.append(int((u, v) in test_edges or (v, u) in test_edges))
        y_score.append(score)

# 计算AUC
auc = roc_auc_score(y_true, y_score)
print(f"AUC: {auc}")

# 计算精确度、召回率和F1分数
precision, recall, f1, _ = precision_recall_fscore_support(y_true, np.array(y_score) > np.median(y_score), average='binary')
print(f"Precision: {precision}")
print(f"Recall: {recall}")
print(f"F1 Score: {f1}")