A Method for Extracting Hypernymy Relationship of Domain Concepts Based on Similar Concept Pairs

Abstract

The invention provides a domain concept hypernym-hyponym relation extraction method based on similar concept pairs. The method comprises the following steps that 1, a domain concept set is composed of multiple domain concepts, and similar domain concepts are extracted from the domain concept set on the basis of a concept clustering method; 2, candidate concept pairs possibly with a hypernym-hyponym relation are obtained, and similar candidate concept pairs are generated according to the similar concepts obtained in the first step; 3, part of training data is obtained by means of a knowledge base, relation characteristics are represented through the similar candidate concept pairs jointly, relation extraction based on multi-sentence characteristics is achieved, and therefore the domain concept hypernym-hyponym relation is extracted. The domain concept hypernym-hyponym relation extraction method based on the similar concept pairs has the advantages that limitation of the corpus scale can be broken through, and the accuracy rate of domain concept hypernym-hyponym relation extraction can be increased by means of the domain concept hypernym-hyponym relation extracted through the multi-sentence characteristics.

Description

technical field [0001] The invention belongs to the technical field of extracting the hyponym relations of domain concepts, and in particular relates to a method for extracting the hyponym relations of domain concepts based on similar concept pairs. Background technique [0002] The semantic understanding of domain concepts and domain concept relationships is a comprehensive application of natural language processing related technologies, and it is also a difficult problem in this field. Domain concept relations mainly include hyponymy relations, synonymous relations, whole-part relations and so on. Among them, the hyponym relationship is a basic semantic relationship, which is the basis for constructing hierarchical ontology. The hyponym relationship refers to the relationship between the hypernym and the hyponym, where the hypernym is a subject term with a wider conceptual extension, and the hyponym is a subject term with a narrower conceptual connotation. For example, f...

AI Technical Summary

Problems solved by technology

This method has a good effect on English with a simple language structure, but there are still several problems: 1) the template needs to be manually formulated, which is time-consuming and labor-intensive; 2) it is difficult to find a completely accurate template in Chinese processing; 3) The recall rate of template extraction is low; 4) Due to the ambiguity of domain concepts, semantic drift is prone to occur in multiple iterations of template extraction, resulting in the introduction of a large number of wrong data and a sharp drop in accuracy

This method only needs a small amount of data annotation, and the extraction effect is better in the combination of small-scale and large-scale data. However, due to errors in template screening, semantic drift will result after multiple iterations.

[0008] (3) Weakly supervised relationship extraction method

This method can obtain more training data and greatly improve the recall rate. However, in a corpus with limited data size and specific domains, due to the lack of training data, the characteristics of sentences cannot be well utilized.

[0009] (4) Unsupervised relationship extraction method

In this case, the weakly supervised relationship extraction method based on the co-occurrence sentence set degenerates into a relationship extraction method based on single sentence level features, resulting in low accuracy of the extraction results

Images