复杂数据上的实体识别综述

A Survey of Entity Recognition on Complex Data

复杂数据的数据量大和数据源不同的特征导致在挖掘复杂数据中的潜在价值时,需要利用实体识别技术。实体识别技术能实现对传统数据进行完整刻画、对数据质量进行管理的重要操作。而在复杂数据进行实体识别具有识别效果差、识别精度不高等问题。本文首先从应用领域的角度探讨复杂数据上的实体识别技术,包括社交网络领域的敏感实体识别、军事领域的目标实体识别、商业领域的商情实体识别。其次,对不同领域中的各个实体识别常用方法进行对比,分析了各个方法的问题与不足。最后,对在不同领域中进行实体识别的难点进行总结。

Complex data is characterized by a large amount of data and different data sources, which lead to the use of entity recognition technology in mining the potential value of complex data. Entity recognition technology can realize some important operations, such as complete description of traditional data and data quality management. However, entity recognition technology applied in complex data has the problems of poor recognition effect and low recognition accuracy. This paper first discusses entity recognition technology on complex data from the perspective of application field, including sensitive entity recognition in social network field, target entity recognition in military field and business entity recognition in commercial field. Secondly, the usual methods of entity recognition in different fields are compared, and the problems and shortcomings of each method are analyzed. Finally, the difficulties of entity recognition in different fields are summarized.