基于对比学习模型的北京独角兽企业关键特征画像分析
DOI:
CSTR:
作者:
作者单位:

1.北京工业大学;2.北京林业大学

作者简介:

通讯作者:

中图分类号:

F279.23,G644

基金项目:

国家自然科学基金项目“科技关联视角下新兴技术弱信号扫描预判方法研究”(项目编号:72074014);国家自然科学基金项目“基于全文本的微观实体扩散机制研究”(项目编号:72004012)


Profiling Key Characteristics of Unicorn Companie in Beijing on the basis of Contrastive Learning Model
Author:
Affiliation:

Fund Project:

  • 摘要
  • |
  • 图/表
  • |
  • 访问统计
  • |
  • 参考文献
  • |
  • 相似文献
  • |
  • 引证文献
  • |
  • 资源附件
  • |
  • 文章评论
    摘要:

    独角兽企业作为科技创新的领跑者和经济增长的风向标,为中国经济发展与科技创新注入巨大活力。大力促进独角兽企业培育与发展,“为创新者画像、为领跑者导航”成为加强创新主体建设的必要举措,因此亟需开展北京独角兽企业画像研究。以170家北京市独角兽企业作为研究对象,以及11 385家瞪羚企业作为对比对象,按照全面性、相对性、高成长特征原则构建北京独角兽企业画像标签,从企查查、智慧芽、政府网站及企业网站上采集有关数据,通过Gower距离计算、最优特征分箱以及对比学习模型识别实现企业关键特征挖掘,最后采用词云技术绘制北京独角兽企业的关键特征画像。研究发现:北京独角兽企业主要为大中型企业,多数分布于海淀区,为科学研究和技术服务业;北京独角兽企业融资表现优异、对外投资活跃,发明专利技术创新水平较高,但专利国际化水平有待提升,同时企业高风险问题值得警惕。为此,独角兽企业需要注意资金的合理利用,并通过确保投资的准确性和长期可持续性进一步提升对外投资活跃度,更重要的是建立健全的法律团队,制定明确的法律合规的企业发展政策,避免潜在的法律纠纷。

    Abstract:

    Unicorn companies, as technology innovators and economic growth indicators, have infused significant dynamism into the city's economic development and technological innovation. This study focuses on unicorn companies in Beijing, with gazelle companies as a contrastive reference. First, the study constructs a profile label system for unicorns using the principles of comprehensiveness, relativity and high-growth characteristics. Next, key characteristics of companies are identified through Gower distance calculation, optimal feature binning and contrastive learning model. Finally, the key characteristics of unicorn companies in Beijing are visualized with word cloud technology. Specifically, a total of 170 Unicorn companies and 11,385 Gazelle companies were collected, and the Unicorn company portrait labels with 30 labels were constructed from 8 dimensions. Subsequently, on the basis of the initial sample set, the Gower distance was used to select the gazelle company comparison set for the Unicorn company experimental set, and the numerical features were discretized based on the optimal information value, and then a total of 4 clusters of Unicorn company's personality characteristics, 10 clusters of gazelle company's personality characteristics, and 5 clusters of common characteristics of the companies were identified using the common and distinctive topic model. Finally, the word cloud map was used to draw 7 company key characteristics portraits. The study indicates that unicorn companies in Beijing exhibit excellent financing performance, active outbound investment, a high level of technological innovation in patented inventions, but the international patents should be improved. In addition, high-risk issues concerning these companies warrant attention. To this end, Unicorn companies need to pay attention to the rational utilisation of funds and further enhance their outbound investment activity by ensuring the accuracy and long-term sustainability of their investments, and more importantly, establish a sound legal team and formulate a clear and legally compliant corporate development policy to avoid potential legal disputes. The findings provide valuable insights to the cultivation and development of unicorn companies and contribute to the economic growth of Beijing.

    参考文献
    相似文献
    引证文献
引用本文

徐硕,马心怡,安欣.基于对比学习模型的北京独角兽企业关键特征画像分析[J].,2024,44(17).

复制
分享
文章指标
  • 点击次数:
  • 下载次数:
  • HTML阅读次数:
  • 引用次数:
历史
  • 收稿日期:2023-12-13
  • 最后修改日期:2024-09-04
  • 录用日期:2024-03-04
  • 在线发布日期: 2025-03-19
  • 出版日期:
文章二维码

联系电话:020-37635126(一、三、五)/83568469(二、四)(查稿)、37674300/82648174(编校)、37635521/82640284(财务)、83549092(传真)

联系地址:广东省广州市先烈中路100号大院60栋3楼302室(510070) 广东省广州市越秀区东风西路207-213星河亚洲金融中心A座8楼(510033)

邮箱:kjgl83568469@126.com kjgl@chinajournal.net.cn

科技管理研究 ® 2025 版权所有
技术支持:北京勤云科技发展有限公司
关闭
关闭