医学实体识别(NER)训练流程/医学关系识别(RE)训练流程

知识图谱 知识抽取的主流流程

1. 数据获取与预处理 (Data Acquisition and Preprocessing)

   • 网络爬虫采集数据 (Web crawling)
   • 数据清洗 (Data cleaning)
   • 文本分词与标准化 (Text tokenization and normalization)

2. 实体识别 (Named Entity Recognition, NER)

   • 识别文本中的命名实体 (Identifying named entities in text)
   • 实体边界确定 (Entity boundary determination)
   • 实体类型分类 (Entity type classification)

3. 关系抽取 (Relation Extraction)

   • 确定实体间的语义关系 (Determining semantic relationships between entities)
   • 关系分类 (Relationship classification)
   • 开放域关系抽取 (Open domain relation extraction)

4. 属性抽取 (Attribute Extraction)

   • 提取实体的属性信息 (Extracting attribute information of entities)
   • 属性值规范化 (Normalizing attribute values)

5. 事件抽取 (Event Extraction)

   • 识别事件触发词 (Identifying even