面向企业模糊型技术需求的信息融合方法研究
Research on Information Fusion Method for Enterprise Fuzzy Technical Requirements
摘要: [目的/意义]：实现需求与供给的精准匹配是产教融合的关键所在，对科技成果转化、技术创新等具有重要意义。然而，企业在发布技术需求时，因商业机密保护等原因，导致技术需求文本描述常具有模糊性，难以与技术供给进行精准匹配。[方法/过程]：设计了面向企业模糊型技术需求的两阶段模型，首先使用BERT BiLSTM CRF算法提取企业发布项目中的需求实体，然后通过挖掘企业基本信息，结合需求实体作为检索条件爬取相关专利成果，通过TF-IDF算法获取专利中关键词作为补充信息，精准化技术需求描述，并识别出技术的聚焦方向与发展前沿。[结果/结论]：以机械制造领域内某企业真实需求作为案例，展示了该方法的完整处理过程，挖掘与TBM相关的施工管理、工件改进、系统研发等关键技术点；通过跨时间维度识别技术需求聚焦方向，TBM与人工智能、大数据、机器学习、探测等技术相结合将是未来发展的趋势。
Abstract: [Purpose/Significance]: Achieving the accurate matching of demand and supply is the key to the integration of industry and education, which is of great significance to the transformation of scientific and technological achievements and technological innovation. However, when enterprises release technical requirements, due to trade secret protection and other reasons, the text description of technical requirements is often ambiguous, and it is difficult to accurately match the technical supply. [Method/Process]: A two-stage model for the fuzzy technical needs of enterprises was designed. Firstly, the BERT BiLSTM CRF algorithm is used to extract the demand entities in the enterprise release projects, and then the relevant patent achievements are crawled by mining the basic information of the enterprise, combined with the demand entities as the search conditions, and the keywords in the patents are obtained as supplementary information through the TF-IDF algorithm, so as to accurately describe the technical requirements and identify the focus direction and development frontier of the technology. [Result/Conclusion]: Taking the real needs of an enterprise in the field of machinery manufacturing as a case, the complete processing process of the method is demonstrated, and the key technical points related to TBM such as construction management, workpiece improvement, and system research and development are excavated. By identifying the focus direction of technology demand across time dimensions, the combination of TBM with artificial intelligence, big data, machine learning, detection and other technologies would be the trend of future development.
文章引用：陶泽奎, 张志清. 面向企业模糊型技术需求的信息融合方法研究[J]. 运筹与模糊学, 2024, 14(2): 857-869. https://doi.org/10.12677/orf.2024.142186

参考文献

[1] 史敏, 罗建, 周斌. 基于竞争情报的企业技术需求识别MTS方法的研究与应用[J]. 图书情报知识, 2018(3): 95-102.
[2] 杨德林, 夏青青, 马晨光. 在线技术转移平台的供需匹配效率分析[J]. 管理科学, 2017, 30(6): 104-112.
[3] Chen, W.L., Zhang, Y.J. and Hitoshi, I. (2006) Chinese Named Entity Recognition with Conditional Random Fields. Proceeding of the 15th Sighan Workshop on Chinese Language Processing, Sydney, Australia, July 2006, 118-121.
[4] Soomro, P.D., Kumar, S., Banbhrani, Shaikh, A.A. and Raj, H. (2017) Bio-NER: Biomedical Named Entity Recognition Using Rule-Based and Statistical Learners. International Journal of Advanced Computer Science and Applications, 8, 163-170. [Google Scholar] [CrossRef] 
[5] 包振山, 宋秉彦, 张文博, 等. 基于半监督学习和规则相结合的中医古籍命名实体识别研究[J]. 中文信息学报, 2022, 36(6): 90-100.
[6] Ma, S.C., Xu, J.H. and Fan, Y. (2022) Characteristics and Key Trends of Global Electric Vehicle Technology Development: A Multi-Method Patent Analysis. Journal of Cleaner Production, 338, 1-15. [Google Scholar] [CrossRef] 
[7] Cao, Q., Cai, H.J., Wang, J.J., et al. (2021) A Scientometric Study of Technological Trend Based on Patent. Journal of Integrated Design and Process Science, 23, 5-28. [Google Scholar] [CrossRef] 
[8] 王京安, 校姜文, 牛建, 汤月. 基于专利分析的技术发展趋势预测研究——以液晶材料技术为例[J]. 科技管理研究, 2019, 39(8): 141-149.
[9] 罗建, 蔡丽君, 史敏. 基于专利的两阶段新兴技术识别研究——以图像识别技术为例[J]. 情报科学, 2019, 37(12): 57-62.
[10] Sarkar, K. and Shaw, K.S. (2017) A Memory-Based Learning Approach for Named Entity Recognition in Hindi. Journal of Intelligent Systems, 26, 301-321. [Google Scholar] [CrossRef] 
[11] 黄水清, 王东波, 何琳. 基于先秦语料库的古汉语地名自动识别模型构建研究[J]. 图书情报工作, 2015, 59(12): 135-140.
[12] Nie, H. (2012) Person-Specific Named Entity Recognition Using SVM with Rich Feature Sets. Chinese Journal of Library and Information Science, 5, 27-46.
[13] Zhang, S., Zhu, H., Xu, H., et al. (2022) A Named Entity Recognition Method towards Product Reviews Based on BiLSTM-Attention-CRF. International Journal of Computational Science and Engineering, 25, 479-489. [Google Scholar] [CrossRef] 
[14] 张召武, 徐彬, 高克宁, 等. 面向教育领域的基于SVR-BiGRU-CRF中文命名实体识别方法[J]. 中文信息学报, 2022, 36(7): 114-122.
[15] Peng, N. and Dredze, M. (2016) Improving Named Entity Recognition for Chinese Social Media with Word Segmentation Representation Learning. Proceedings of the 54th Annual Meeting of the Association for Computational Linguistice, Berlin, Germany, August 2016, 149-155. [Google Scholar] [CrossRef] 
[16] Wu, F.Z., Liu, J.X., Wu, C.H., et al. (2019) Neural Chinese Named Entity Recognition via CNN-LSTM-CRF and Joint Training with Word Segmentation. Proceedings of the World Wide Web Conference, New York NY, 13-17 May 2019, 3342-3348. [Google Scholar] [CrossRef] 
[17] 翟东升, 李梦洋, 何喜军, 徐硕. 非线性技术演化条件下的专利研发投资决策研究[J]. 中国管理科学, 2021, 29(12): 168-178.
[18] Devlin, J., Chang, M.W., Lee, K., et al. (2019) BERT: Pre-Training of Deep Bidirectional Transformers for Language Understanding. 4171-4186.
https://arxiv.org/abs/1810.04805
[19] Graves, A. and Schmidhuber, J. (2005) Framewise Phoneme Classification with Bidirectional LSTM and Other Neural Network Architectures. Neural Network, 18, 602-610. [Google Scholar] [CrossRef] [PubMed]

为你推荐