基于GRU-Attention机制的高校发明专利质量评估模型研究

Research on the Quality Evaluation Model of University Invention Patent Based on GRU-Attention Mechanism

摘要:

【目的/意义】近年来中国的专利数量增速惊人，高校作为专利的重要产出地之一，其发明专利的申请数量也与日俱增，但高校发明专利的质量情况却不尽如人意。深度学习技术的飞速发展为专利质量评估提供了一种新的思路。【方法/过程】为了能够快速准确地对高校发明专利质量进行评估，本文充分考虑专利的文本属性和结构化数据信息，基于深度学习技术构建了基于GRU-Attention机制的高校发明专利质量评估模型，并通过引入衡量高校发明专利质量的关键指标及权重信息对初始模型进行改进。【结果/结论】经验证，结合专家知识改进后的评估模型在稳定性和准确率上有明显优势，能够为高校发明专利质量评估提供了有效的理论和实践支撑。

Abstract:

In recent years, the number of patents in China has increased at an alarming rate. As one of the important production areas of patents, the number of applications for invention patents in universities is also increasing day by day, but the quality of invention patents in universities is not satisfactory. The rapid development of deep learning technology provides a new idea for the evaluation of patent quality. In order to be able to quickly and accurately evaluate the quality of university invention patents, this paper takes full account of the text attribute and structured data information of patents, constructs a university invention patent quality evaluation model based on Gru-attention mechanism based on deep learning technology, and improves the initial model by introducing the key indicators and weight information to measure the quality of university invention patents. It has been verified that the improved evaluation model combined with expert knowledge has obvious advantages in stability and accuracy, which can provide effective theoretical and practical support for the quality evaluation of invention patents in universities.