A Novel Entity Relation Extraction Approach Based on Micro-Blog
Entity relation extraction is a key task in information extraction. The purpose is to find out the semantic relation between entities in the text. An improved tree kernel-based method for relation extraction described in this paper adds the predicate verb
- PDF / 933,384 Bytes
- 10 Pages / 439.37 x 666.142 pts Page_size
- 33 Downloads / 231 Views
Abstract. Entity relation extraction is a key task in information extraction. The purpose is to find out the semantic relation between entities in the text. An improved tree kernel-based method for relation extraction described in this paper adds the predicate verb information associated with entity, prunes the original parse tree, and removes some redundant structure on the basis of the Path-enclosed Tree. The experiment shows that the proposed method delivered better performance than existing methods. Keywords: Relation extraction Tree kernel-based method Convolution tree kernel
1 Introduction With the rapid development of science and technology, especially internet technology, the rapidly increasing network information in the real world is far beyond the ability of human reading. The difficulty people face is how to filter the useless information and how to draw out specific information that people need. Information extraction is originally presented at the Message Understanding Conferences (MUC, 1987–1998) by the Defense Advanced Research Projects Agency (DARPA), and seen as an important branch in the field of Natural Language Processing (NLP). After the MUC suspended, the Automatic Content Extraction program (ACE) [1] supported by National Institute of Standards and Technology (NIST) aims at further promoting the development of information extraction. The application scope of relation extraction is becoming more and more wide. Relation extraction has played an important role in the development of information extraction, question answering system, machine translation and so on, and its purpose is to find out semantic relations between entities from the tagged text. In this paper, we use the tree kernel-based method [2–6] for relation extraction, which calculates the similarity by counting the number of the same sub-structures between two relation instances. As the tree kernel-based method can make full use of
Project supported by the National Nature Science Foundation of China (No. 61271413, 61472329, 61532009), Innovation Fund of Postgraduate, Xihua University. © Springer International Publishing Switzerland 2016 D.-S. Huang et al. (Eds.): ICIC 2016, Part III, LNAI 9773, pp. 401–410, 2016. DOI: 10.1007/978-3-319-42297-8_38
402
H. Zheng et al.
the structural information which can’t be expressed by the feature vector-based methods, more and more researchers begin to study and use it in recent years. In this paper, the improved tree kernel-based method we propose reduces the redundant information and expands the original tree structure, so that it can contain more semantic information. Experiment shows that the proposed method can significantly improve the performance of the relation extraction. This paper is organized as follows. In Sect. 2, we review previous work on relation extraction. In Sect. 3, we present our tree kernel-based methods to relation extraction. In Sect. 4, we analyze the results of our experiments. Finally, we present conclusions in Sect. 5.
2 Previous Works Previous approaches to relatio
Data Loading...