Abstract:The grade of patent quality can be used as a reference for the patentee to make decisions about maintaining, abandoning, or converting a patent. The patent quality evaluation process, combined with subjective scoring by experts and quantitative analysis of objective models, can improve the scientificity and interpretability of the evaluation results. Therefore, based on the perspective of patent users, through expert consultation method to construct a multi-dimensional patent quality evaluation index system that includes technical, legal, market, and subject characteristics. Evaluate patents in the field of machine learning technology, Using a combination weighting method combining analytic hierarchy process and entropy weight method to calculate the weight of feature indicators, while introducing matter element analysis theory, Calculate the maximum membership degree of each index of the patent to be evaluated for the quality level to evaluate the quality of the patent, and select a patent related to electrical data digital processing technology for evaluation and measurement. The results show that the value of the patent in the technical, legal, and subject dimensions is Level I, the value in the market dimension is Level III, and the comprehensive quality is Level I. The patent quality grade obtained from the evaluation is consistent with the actual quality results, which proves that the evaluation model based on combination weighting and cloud matter-element has certain feasibility and applicability in patent quality evaluation. Subsequent research can use this model to evaluate patent portfolios.