Semi-supervised entity relation extraction based on trigger word
Liting Tai, Fen-Zhuo Guo, Su‐Juan Qin · 2017
Relation extraction is one of the most important tasks in the process of information extraction. This paper studies a semi-supervised entity relation extraction method, most of the existing semi-supervised relation extraction methods are based on the bootstrapping algorithm, the obvious drawback of this method is that it is prone to semantic drift, resulting in the lower recall rate. In order to solve this problem, this paper based on the traditional method of trigger words extraction, presents an advanced trigger words extraction method (ATW), it is more flexible and accurate, effectively alleviates the problem of semantic drift. The experimental result show that the method has a good performance, precision, recall rate and F value have been greatly improved.