Patent valuation method and valuation system based on data mining and heterogeneous knowledge association

Abstract

The invention belongs to the technical field of patent value evaluation, and discloses a patent valuation method and system based on data mining and heterogeneous knowledge association, and the methodcomprises the steps: constructing a heterogeneous knowledge association network as a patent valuation network environment according to the complex association relation between a patent text and a market information text; performing accurate measurement on patent values in the network environment by training a machine learning and data mining model; and in the network environment, accurately measuring the patent value through machine learning and data mining methods. Related value features are extracted from patents, and complex external entity features are introduced from incidence relation analysis. A huge heterogeneous knowledge association network is formed by establishing association between a large number of patent texts and complex texts related to an external market, and a machinelearning and data mining model is trained through extracted value features of patents themselves and exteriors to perform accurate measurement on patent values.

Description

technical field [0001] The invention belongs to the technical field of patent value evaluation, and in particular relates to a patent evaluation method and a patent evaluation system based on data mining and heterogeneous knowledge association. Background technique [0002] Currently, the closest prior art: [0003] At present, the common value measurement models mainly include the method of counting the number of citations, the method of latent graph model, the method of deep learning, etc. These methods use the value characteristics of the patent itself as the starting point or introduce simple external characteristics to study the value of patents. [0004] (1) From the perspective of the patent itself, the current method believes that the value of the patent is single and definite. [0005] For intangible assets such as patents, their value itself changes with changes in external factors, and cannot always be a stable value. Even in different environments, different va...

AI Technical Summary

Problems solved by technology

However, the existing methods look at the value of patents from a single and definite perspective, which will bury the value of patents from other perspectives, thus making it impossible to make accurate valuations of patents

[0006] (2) From the perspective of correlation, the current method believes that the value of patents is non-correlated and non-market

The value of patents is closely related to market conditions, but how to subtly associate patents with market conditions and how to convert them into computer-recognizable features is a technical difficulty

Images