欢迎您！东篱公司
退出
-
申报数据库
-
立项数据库
-
成果数据库
-
项目获奖数据库
中文摘要：
利用关联实体识别技术可以对异构网络中主题相关的实体进行检测并整合，更好地帮助用户理解搜索目标.然而，目前关联实体识别技术考虑的因素较为单一、对识别结果缺乏验证而影响识别准确性.文中提出了一种两阶段的关联实体识别模型，充分考虑了实体的模式特征与属性特征.此外，提出了一种增量式验证算法，基于迭代对识别结果进行增量式的验证并修正，以保证结果的准确性.通过实验验证了文中所提出的关键技术的可行性和有效性.
英文摘要：
Related entity identification is a necessary technique to find and integrate the entities that are related tightly in heterogeneous networks. It is useful to make users understand the search result better. However, eurrent techniques consider limited influence factors for related entity identification and lack verification which often lead to dumb results. In this paper, we present a two-phase related entity identification model by fully considering entity schema and entity feature. Also an incremental verification algorithm is proposed to iteratively verify and repair the result of identification. The experiments demonstrate the feasibility and effectiveness of our key techniques.
同期刊论文项目
同项目期刊论文
期刊信息
位置：