This study aims to construct a patent value evaluation model that can batch identify and filter patents with market transformation potential,which improves patent management efficiency and promotes the market application of patent technology.The study employs multivariate statistical analysis and artificial neural network methods,including preliminary screening of key indicators and dimensionality reduction using principal component analysis,ultimately constructing a multi-layer perceptron(MLP)model with an input layer,two hidden layers,and an output layer.The model is optimized for recall and precision using ten-fold cross-validation.The results indicate that the model can effectively predict the value of patents.Through empirical testing of"20+8"industries patents in Shenzhen,it was found that tested patents generally have a higher value,with A-level and above patents accounting for 11.27%,and B-level and above patents accounting for 41.82%.
NETL
NSTL
万方数据
