The quantity and quality of patents have become an important indicator to measure the competitiveness of science and technology of a country or industry. Although the number of patent applications in China has been growing rapidly, there are stillroom for improvementinterms ofthe quality of patents.The team would utilizeNaturalLanguageProcessing based onDeepLearning to construct the correlation network between patents, so that the similarity measurement of patent document not only takes the patent citation information into account, but also considers the context semantics in the documenttomake a more accurate quantitative evaluation of the influence and technology frontier of each patent. On this basis, we can evaluate the patent quality of different technology segments and different enterprises or regions, identify core patents, and analyze the evolution process and development path of technology. This is of great guiding significance to enhance the national innovation ability and improve the technological competitive advantage of enterprises.
