一种基于相似度计算的实体数据关系归属方法
A Relationship Attribution Method for Entity Data Based on Similarity Calculation

特定领域的数据蕴含了大量有价值的知识及关系划分，从中能正确将其进行关系划分一直是一个值得关注的话题。当前关系划分都依赖于大量样本模型进行训练得出，由于特定领域的实体数据关系样本数量较少，显然应用到特定领域中存在局限。因此本文针对该问题，提出一种基于相似度计算的实体数据关系归属方法，其中建立一个特定领域的少样本实体关系术语树，与待划分的实体数据进行相似度计算得到在树中具体位置，从而解决错误归属问题，显著减少人工管理成本，能够有效提升系统的可用性。
Domain-specific data contains a large amount of valuable knowledge and relationship delineation, from which it is always a topic of interest to be able to correctly perform relationship delineation. Currently, the relationship classification relies on a large number of sample models for training, due to the small number of domain-specific entity data relationship samples, it is obvious that there are limitations in applying to specific domains. Therefore, in this paper, we propose a similarity-based relationship attribution method for entity data, in which a domain-specific entity relationship term tree with few samples is established, and the entity data to be partitioned is similarity-calculated to get the specific position in the tree, thus solving the problem of misattribution, significantly reducing the cost of manual management, and effectively improving the usability of the system.