图书情报工作

图书情报工作 >

2026 , Vol. 70 >Issue 9: 27 - 41

DOI: https://doi.org/10.13266/j.issn.0252-3116.2026.09.003

研究论文

基于“指标+语义”特征和SSA算法的潜在高价值专利识别方法

  • 窦路遥 ,
  • 魏凤 ,
  • 周洪 ,
  • 邓阿妹
展开
  • 1 中国科学院武汉文献情报中心 武汉 430071;
    2 中国科学院大学经济与管理学院信息资源管理系 北京 100190;
    3 科技大数据湖北省重点实验室 武汉 430071
窦路遥,博士研究生;魏凤,研究员,博士,通信作者,E-mail:weif@whlib.ac.cn;周洪,副研究员,博士;邓阿妹,副研究员,博士。

收稿日期: 2025-03-11

  修回日期: 2025-08-28

  网络出版日期: 2026-04-30

基金资助

本文系国家电网公司总部科技项目“双碳目标下电力绿色低碳关键支撑技术评价方法和专利标准化研究”(项目编号:1400-202340338A-1-1-ZN)研究成果之一。

收起

Identification of Potential High-Value Patents Based on “Indicator + Semantic” Features and the Sparrow Search Algorithm

  • Dou Luyao ,
  • Wei Feng ,
  • Zhou Hong ,
  • Deng Amei
Expand
  • 1 National Science Library (Wuhan), Chinese Academy of Sciences, Wuhan 430071;
    2 Department of Information Resources Management, School of Economics and Management, University of Chinese Academy of Sciences, Beijing 100190;
    3 Hubei Key Laboratory of Science and Technology Big Data, Wuhan 430071
Zhang Meng, Doctoral Candidate;Mu Dongmei, Professor, PhD, Corresponding Author, E-mail:moudm@jlu.edu.cn;Zhang Ning, Master’s Candidate;Wang Ping, Postdoc;Zhang Shengxia, Business Manager;Wu Fei, Project Manager;Zhang Xingna, Master’s Degree;Liu Lulu, Product Manager.

Received date: 2025-03-11

  Revised date: 2025-08-28

  Online published: 2026-04-30

Supported by

This work is supported by the Science and Technology Project of the Headquarters of State Grid Corporation of China titled “Research on Evaluation Methods of Key Supporting Technologies for Electric Power Green and Low-Carbon Development and Patent Standardization under the ‘Dual Carbon’ Goal” (Grant No. 1400-202340338A-1-1-ZN).

Fold

摘要

[目的/意义] 融合专利价值层面的指标特征和专利文本层面的语义特征，并基于麻雀优化算法(SSA)来解决现阶段潜在高价值专利识别过程中的特征不全面问题和参数组合优化问题。[方法/过程] 提出基于“指标+语义”特征和SSA算法的潜在高价值专利识别模型(SSA-BERT-CNN-BiLSTM, SSA-BCB)。在特征选取上，通过向量拼接融合以BERT模型提取出的专利文本语义特征和以共现网络形式挖掘出的指标特征。在模型优化上，基于SSA算法实现CNN-BiLSTM模型的参数组合优化。[结果/结论] 结果表明，SSA-BCB模型在潜在高价值专利识别任务中具备一定优势，F1值较之各基线模型提升约4.9%,AUC值达0.927；结合注意力机制后，模型的性能普遍提升约4.2%。研究说明SSA算法对于潜在高价值专利识别任务具备提升效果，语义关系的复杂程度也会进一步影响模型识别效果，优化后的参数组合与注意力机制的结合能进一步加强模型识别性能。

关键词： 麻雀优化算法; 语义特征; 潜在高价值专利; SSA-BCB; 指标特征

本文引用格式

窦路遥 , 魏凤 , 周洪 , 邓阿妹 . 基于“指标+语义”特征和SSA算法的潜在高价值专利识别方法[J]. 图书情报工作, 2026 , 70(9) : 27 -41 . DOI: 10.13266/j.issn.0252-3116.2026.09.003

Abstract

[Purpose/Significance] This study integrates indicator features reflecting patent value with semantic features derived from patent texts and leverages the Sparrow Search Algorithm(SSA) to address two major challenges in the identification of potential high-value patents: incomplete feature representation and the complexity of parameter optimization. [Method/Process] It proposes a potential high-value patent identification model, named SSA-BERT-CNN-BiLSTM(SSA-BCB), which incorporates both semantic and indicator features. Semantic features are extracted from patent texts using the BERT model, while indicator features are mined from co-occurrence networks. These heterogeneous features are fused via vector concatenation. The CNN-BiLSTM model parameters are subsequently optimized through SSA to enhance classification performance. [Result/Conclusion] Experimental evaluation demonstrates that the proposed SSA-BCB model achieves a 4.9% improvement in the F1-score over benchmark models, with the area under the ROC curve(AUC) reaching 0.927. Furthermore, the integration of an attention mechanism yields an additional average performance gain of 4.2%. These findings confirm that SSA effectively enhances the identification of potential high-value patents, that the complexity of semantic relationships can substantially influence model performance, and that the combination of optimized parameters with attention mechanisms can further improve identification accuracy.

Key words： sparrow search algorithm; semantic representation; high-value patent identification; SSA-BCB; indicator features

参考文献

[1] 蔡鸿宇,王向锋,徐宗煌,等.高价值专利有何特征?--基于模糊集定性比较分析的组态效应研究[J].情报科学, 2024, 42(4):167-174, 183.(CAI H Y, WANG X F, XU Z H, et al. What are the characteristics of high-value patents?A configuration effect study based on fuzzy-set qualitative comparative analysis[J]. Information science, 2024, 42(4):167-174, 183.)
[2] 李黎明.专利价值研究的文献综述与未来展望[J].情报杂志,2023, 42(2):166-174.(LI L M. A literature review and future outlook on patent value research[J]. Journal of intelligence,2023, 42(2):166-174.)
[3] 付振康,柳炳祥,鄢春根,等.专利寿命视角下潜在高价值专利识别方法研究[J].情报杂志, 2023, 42(5):154-161, 207.(FU Z K, LIU B X, YAN C G, et al. Research on identification methods for potential high-value patents from the perspective of patent lifespan[J]. Journal of intelligence, 2023, 42(5):154-161,207.)
[4] 胡泽文,周西姬,任萍.基于扎根理论的高价值专利评估与识别研究综述[J].情报科学, 2022, 40(2):183-192.(HU Z W,ZHOU X J, REN P. A review of high-value patent evaluation and identification research based on grounded theory[J]. Information science, 2022, 40(2):183-192.)
[5] 李小童,徐菲.高价值专利识别方法有效性实证研究[J].科技与法律, 2019(1):11-17.(LI X T, XU F. Empirical study on the effectiveness of high-value patent identification methods[J].Science and law, 2019(1):11-17.)
[6] 胡泽文,周西姬.基于BP神经网络和MIV算法的高价值专利预测与影响因素分析[J].信息资源管理学报, 2023, 13(6):144-155.(HU Z W, ZHOU X J. High-Value patent prediction and influencing factor analysis based on BP neural network and MIV algorithm[J]. Journal of information resources management,2023, 13(6):144-155.)
[7] 周潇,王博,胡玉琳,等.基于时序图神经网络的潜在高价值专利识别研究[J].情报学报, 2024, 43(6):697-711.(ZHOU X, WANG B, HU Y L, et al. Research on identifying potential high-value patents based on temporal graph neural networks[J]. Journal of the China Society for Scientific and Technical Information, 2024, 43(6):697-711.)
[8] 胡嘉蕴,赵楠.适用于医药生物领域审查阶段的专利价值评估体系的建立与验证[J].中国发明与专利, 2019, 16(9):90-97.(HU J Y, ZHAO N. Establishment and validation of a patent value evaluation system for the pharmaceutical and biotechnology field during the examination stage[J]. China invention&patent, 2019, 16(9):90-97.)
[9] 冉从敬,李旺.校企合作视角下高校高价值专利推荐模型构建--以"新能源汽车"为例[J].图书情报工作, 2023,67(11):115-126.(RAN C J, LI W. Construction of a highvalue patent recommendation model for universities from the perspective of university-industry collaboration:a case study of "new energy vehicles" [J]. Library and information service,2023, 67(11):115-126.)
[10] 洪群业,汪志波,王乐,等.基于熵权法的我国烟草行业高价值专利评价分析[J].烟草科技, 2024, 57(11):106-112.(HONG Q Y, WANG Z B, WANG L, et al. Evaluation and analysis of high-value patent in China's tobacco industry based on the entropy weight method[J]. Tobacco science&technology, 2024,57(11):106-112.)
[11] 许振亮,王振,高天艳.基于层次分析法-德尔菲法的高价值专利筛选研究--以稀土永磁产业为例[J].中国科技论坛, 2024(1):62-71.(XU Z L, WANG Z, GAO T Y. Research on high-value patent screening based on AHP-Delphi method:a case study of the rare earth permanent magnet industry[J]. China science and technology forum, 2024(1):62-71.)
[12] 孔德婧,董放,陈子婧,等.离群专利视角下的新兴技术预测--基于BERT模型和深度神经网络[J].图书情报工作,2021, 65(17):131-141.(KONG D J, DONG F, CHEN Z J, et al.Emerging technology prediction from the perspective of outlier patents:based on BERT model and deep neural networks[J].Library and information service, 2021, 65(17):131-141.)
[13] 赵雪峰,胡瑾瑾,吴德林,等.一种基于特征拼接、标签迁移及深度学习组合的专利价值评估方法[J].情报学报,2023, 42(6):663-680.(ZHAO X F, HU J J, WU D L, et al. A patent value evaluation method based on feature stitching, label migration, and deep learning combination[J]. Journal of the China Society for Scientific and Technical Information, 2023,42(6):663-680.)
[14] 程文银,胡鞍钢,陈雪丽.知识产权强国背景下中国高价值专利发展:测度与实证分析[J].北京工业大学学报(社会科学版), 2022, 22(5):1-12.(CHENG W Y, HU A G, CHEN X L. The development of high-value patents in China under the context of building a strong intellectual property country:measurement and empirical analysis[J]. Journal of Beijing University of Technology (social science edition), 2022, 22(5):1-12.)
[15] 任海英,孙闯闯.融合知识网络嵌入特征的高价值专利预测[J].北京工业大学学报(社会科学版), 2023, 23(5):138-152.(REN H Y, SUN C C. High-Value patent prediction integrating knowledge network embedding features[J]. Journal of Beijing University of Technology (social science edition), 2023, 23(5):138-152.)
[16] LIU W, CHENG C, GAN K, et al. A transfer learning based patent transfer prediction model:discussion about the condition of China and US[J]. Expert systems with applications, 2025, 259:125335.
[17] CHOI J, JEONG B, YOON J, et al. A novel approach to evaluating the business potential of intellectual properties:a machine learning-based predictive analysis of patent lifetime[J].Computers&industrial engineering, 2020, 145(11):106-124.
[18] JIANKAI X, BO S. A survey on sparrow search algorithms and their applications[J]. International journal of systems science,2024, 55(4):814-832.
[19] LI H, WEN H, LI J, et al. SLIC-SSA:an image segmentation method based on superpixel and sparrow search algorithm[J].International journal of computational science and engineering,2024, 27(2):182-194.
[20] ACHEAMPONG F A, NUNOO-MENSAH H, Chen W.Transformer models for text-based emotion detection:a review of BERT-based approaches[J]. Artificial intelligence review,2021, 54(8):5789-5829.
[21] FIRDAUS D, SUMARDI I. Comparative analysis of graph neural network with SAGE conv, GAT conv, and GCN conv techniques for fake news detection[J]. Industrial sciencetech journal, 2024, 1(1):54-61.
[22] CHEN R, WANG Y, LI G, et al. Pre-training models based knowledge graph multi-hop reasoning for smart grid technology[C]//Proceedings of 2021 5th Chinese conference on swarm intelligence and cooperative control. Singapore:Springer Nature, 2022:1866-1875.
[23] 窦路遥,魏凤,邓阿妹,等.基于维度复合和可解释性说明的高价值专利评估体系构建研究[J].现代情报, 2025, 45(11):30-42.(DOU L Y, WEI F, DENG A M, et al. Research on the construction of a High-Value Patent evaluation system Based on dimensional composite and interpretability explanation[J].Journal of modern information, 2025, 45(11):30-42.)
[24] 李跃艳,熊回香,李晓敏.基于主成分分析法的期刊评价模型构建[J].情报杂志, 2019, 38(7):199-207.(LI Y Y, XIONG H X, LI X M. Construction of journal evaluation model based on principal component analysis[J]. Journal of intelligence, 2019,38(7):199-207.)
[25] XIAO L, WANG G, ZUO Y. Research on patent text classification based on word2vec and LSTM[C]//2018 11th international symposium on computational intelligence and design. Ningbo:IEEE, 2018:71-74.
[26] FAIRUZABADI M, KUSRINI K, UTAMI E, et al.Advancements and challenges in gated recurrent units (GRU) for text classification:a systematic literature review[C]//20247th International conference of computer and informatics engineering. Indonesia:IEEE, 2024:1-7.
[27] ZHANG Y, ZHANG S, MU J. A patent quality evaluation model based on ALBERT and BiGRU[C]//Third international conference on electronic information engineering, big data, and computer technology. United States:SPIE, 2024:2135-2139.
[28] ZHANG Y, WANG Q. Patent prediction based on long short-term memory recurrent neural network[C]//Proceedings of the 9th International conference on computer engineering and networks.Singapore:Springer Nature, 2021:291-299.
[29] LEE A, CHOI K, KIM G. LDA topic modeling and recommendation of similar patent document using word2vec[J].Information systems review, 2020, 22(1):17-31.
[30] LYU L, HAN T. A comparative study of Chinese patent literature automatic classification based on deep learning[C]//2019 ACM/IEEE joint conference on digital libraries. United States:IEEE,2019:345-346.
[31] 张彪,吴红,高道斌,等.基于潜在高被引论文与高价值专利的创新前沿识别研究[J].图书情报工作, 2022, 66(18):72-83.(ZHANG B, WU H, GAO D B, et al. Research on identification of innovation frontiers based on potential highly cited papers and high-value patent[J]. Library and information service, 2022,66(18):72-83.)
[32] 丁倩.基于多层遴选的农业专利价值评估方法研究[D].北京:中国农业科学院, 2024.(DING Q. Research on the value evaluation method of agricultural patents based on multilayer selection[D]. Beijing:Chinese Academy of Agricultural Sciences, 2024.)
[33] JIANG X, ZHOU P, ZHAO S, et al. Attention mechanism based technical phrase extraction for patent text[C]//2023 8th International conference on intelligent computing and signal processing. Xi'an, China:IEEE, 2023:1276-1279.
Options
文章导航

/

〈 〉