Research On Application Of BP Neural Network In Valuation Of Patent

With the growing of social and economic development, people’s growingawareness of knowledge and patent is more and more obviously. More and moreattention on the importance of patents is paid by people. As a special intangible asset,the patent shows a distinct specificity of value, which results at the particular reflectionof the valuation of the patent itself. This assessment of the value to the patent becomesmore difficult. How to evaluate the value of patents effectively becomes an importanttopic of current research.However, the researches show a lack of considerations of patent value assessmentfactors that affect the value of patents. This paper aims to comprehensively examine thefactors that affect the value of patents, extract relevant factor index, and build acomprehensive scientific assessment of the value of patents BP neural network modelwith BP neural network nonlinear high-speed processing capability.This neural network by constructing patent value-oriented evaluation system toensure comprehensive and representative indicators, and indicators, a series ofstandardized, making the index data to better adapt to BP neural network. By collectingexisting data samples, and samples for scientific data processing, includingnormalization, principal component analysis to reduce the dimensions of such treatment,and then use the training samples for BP neural network training and correction oferrors to achieve the desired accuracy, and then the test sample test corrected BP neuralnetwork to predict test samples patent value and the actual value of patents through acomparative analysis, which proves the value of BP neural network patented thefeasibility and effectiveness evaluation model for the future patent valuation reliablyprovides a fast, comprehensive and scientific evaluation.