An Empirical Study on the Four-Quadrant Classification of University Patents:Taking Transformation Probability and Transformation Amount Expectation as Dimensions
[Purpose/significance]Making a classification evaluation of university patents and taking targeted disposal strategies can optimize resource allocation,avoid resource wastage caused by indiscriminate management,and improve the efficiency of patent transformation.[Method/process]Taking the real patent transformation data from 16 universities as research objects,four machine learning algorithms are compared to establish patent transformation probability classification prediction model and transformation amount classification prediction model.Based on these two models,a four-quadrant classification model for university patents is constructed,and empirical comparative analysis is conducted.[Result/conclusion]When classifying and evaluating the patent transformability and transformation amount in universities,the machine learning model based on random forest algorithm is more effective.The proportion of problematic patents in universities is generally higher.Different and targeted classification and disposal strategies can be adopted for pa-tents in different quadrants.
万方数据
