漫谈·知识图谱

1、关系抽取

1.1 专家系统

• 以两家公司有 投资 关系为例

key_words = ["收购","竞拍","转让","扩张","并购","注资","整合","并入","竞购","竞买","支付","收购价","收购价格","承购","购得","购进",
             "购入","买进","买入","赎买","购销","议购","函购","函售","抛售","售卖","销售","转售"]

1.2 句法依存

• 根据人的说话习惯来抽取
• 主谓宾

if 'SBV' in child_dict and 'VOB' in child_dict:
	r = words[index]
	e1 = complete_e(words, postags, child_dict_list, child_dict['SBV'][0])
	e2 = complete_e(words, postags, child_dict_list, child_dict['VOB'][0])
	svos.append([e1, r, e2])

1.3 远程监督

2、知识图谱的常见表示（Neo4j）

2.1 数据

字段	描述
Person_a	实体a（人名，如：贾宝玉）
Person_b	实体b（人名，如：薛宝钗）
Relation	实体间关系（丈夫）
Family_a	实体a的家族（贾家荣国府）
Family_b	实体b的家族（薛家）

2.2 导入

with open("./raw_data/relation.txt") as f:
    for line in f.readlines():
        rela_array=line.strip("\n").split(",")
        print(rela_array)
        graph.run("MERGE(p: Person{cate:'%s',Name: '%s'})"%(rela_array[3],rela_array[0]))
        graph.run("MERGE(p: Person{cate:'%s',Name: '%s'})" % (rela_array[4], rela_array[1]))
        graph.run(
            "MATCH(e: Person), (cc: Person) \
            WHERE e.Name='%s' AND cc.Name='%s'\
            CREATE(e)-[r:%s{relation: '%s'}]->(cc)\
            RETURN r" % (rela_array[0], rela_array[1], rela_array[2], rela_array[2])
        )

2.3 查询

match(p )-[r]->(n:Person{Name:'%s'}) return  p.Name,r.relation,n.Name,p.cate,n.cate Union all match(p:Person {Name:'%s'}) -[r]->(n) return p.Name, r.relation, n.Name, p.cate, n.cate

2.4 效果

• 贾宝玉 的关系图谱
  
• 李纨 的关系图谱
  
• 林黛玉 的关系图谱