Research on patent value evaluation methods for medical universities and colleges
WANG Jian-bo
OUYANG Zhao-lian
CHEN Juan
LU Yan
ZHANG Ting
Abstract：Objective:Given that traditional patent value evaluation methods are easily influenced by subjective factors and struggle to comprehensively and objectively reflect the value of patents,this study proposed a patent value evaluation method to identify high-value patents.Methods:This study employs information gain and variance analysis methods to perform feature selection on patent data from medical universities,optimizing input features and reducing redundant information.In terms of model construction,a three-layer BP neural network structure was designed,incorporating ReLU activation functions and Dropout techniques to mitigate overfitting issues and enhance the model's ability to handle nonlinear relationships,thereby improving the model's stability and generalization.Results:The experimental results showed that the model achieved a prediction accuracy of 96%.Feature selection effectively improved the model's performance to a certain extent,demonstrating good results across multiple evaluation metrics and verifying the feasibility of the proposed method.Conclusion:This study provides a feasible and effective practical method for patent value evaluation in medical universities,offering valuable reference and application significance,particularly in the identification and prediction of high-value patents.
Keywords：patent value evaluationBP neural networkinformation gainvariance analysisfeature selection
Publication Date：2026-06-30
Online Publishing Date：2026-08-14(First online date of this platform, not the publication date of the document)
Pages：8( 1243-1250 )