Comprehensive Evaluation Of Patent Quality Based On Unascertained Clustering

With people's attention to intellectual property,the number of patents as an important representation of intellectual property has exploded,but the quality of patents has not increased.A large number of low-quality patents not only have a limited effect,but also lead to the waste of social resources and stifle innovation.The comprehensive evaluation of patents helps to understand the technical status of the target patents and to screen the core patents.It also enables researchers to understand the current development status of the industry and predict the future development trend.Firstly,this paper analyzes the patent quality indexes at home and abroad,selects the indexes that have great influence on the patent quality,and constructs the patent quality evaluation index model.At the same time,the characteristic type of the data was analyzed by taking the steel industry-related patents as the target data set,and the unascertained clustering algorithm and fuzzy mean clustering algorithm suitable for the data type were selected to analyze and evaluate the target patent quality.Finally,the target patent data is clustered into different levels to obtain high-quality patents.In the setting of patent quality indicators,this paper analyzes the indicators related to patent quality one by one,and finally selects the patent citation information,the number of patent families and the number of rights declarations to construct the patent quality according to the principles of science,validity and feasibility.In the process of clustering,unascertained clustering algorithm is found to have good performance in efficiency and accuracy.This method is based on the analysis of objective data and data mining to obtain the evaluation results,which is different from the grade evaluation of manual judgment and more conducive to computer operation.At the same time,the algorithm can not only figure out which grade the target patent belongs to,but also figure out the probability of the patent belonging to other grades.It is a new exploration based on the application of unascertained theory,expecting to achieve meaningful innovative results in patent analysis and evaluation.