Ontology-based patent technology topic clustering method

Abstract

The invention relates to an ontology-based patent technology topic clustering method. The method comprises the following steps: establishing a mapping relationship between words in patent technology topics and concept terms in a domain ontology; calculating similarities among the concept terms; using the similarities among the concept terms to calculate similarities among the patent technology topics; and inputting the similarities among the patent technology topics into an AP clustering algorithm as a matrix, and outputting a plurality of clusters obtained through gathering of the patent technology topics, and a topic corresponding to each cluster. The ontology-based patent technology topic clustering method provided by the invention has the advantages that the extracted key technology topic words are clustered based on similarity calculation, the words with similar meanings are gathered into a cluster, and the word which is the most representative of each cluster is selected as the technology topic of the cluster, so that the accuracy, a recall rate and comprehensive evaluation index values of clustering results are improved, and requirements of practical use can be greatly satisfied.

Description

technical field [0001] The invention belongs to the technical field of text information processing, and in particular relates to an ontology-based clustering method for patent technology topics. Background technique [0002] Intellectual property is an important means of modern production protection and promotion of scientific and technological progress. In the context of globalization, the competition of intellectual property rights determines the future competition in the international market. Intellectual property rights represent the technological status and core competitiveness of countries and enterprises. As an important carrier of intellectual property rights, patent documents contain a large amount of technical, commercial and legal information, and are a rich mine of knowledge mining. In today's increasingly fierce international competition, various countries and regions continue to accelerate the pace of informationization and intelligent society, so the signifi...

AI Technical Summary

Problems solved by technology

However, in the process of constructing the patented technology efficacy matrix, if the extracted key technologies are directly used to construct the technical efficacy matrix, it is easy to cause the problem of large matrix dimension and sparse data

Therefore, it is necessary to perform clustering operations on the extracted technical topics. In terms of word clustering, the defects of the existing technology are: the similarity calculation based on word form cannot achieve good results on Chinese corpus, so it is not applicable In the extraction of technical topics in patents; when calculating the complexity of words, the method of calculating the similarity of words based on the degree of co-occurrence of words in two words is too rough and does not consider the semantic information between words

The above defects lead to the low accuracy rate, recall rate and comprehensive evaluation index value of the patent technology subject clustering results of the prior art

Images