Scholarlink

<p>Nowadays, innovation is a key driver of the economy and the competitiveness of organizations, helping them to survive in a global market. The ability to anticipate innovations in the market is thus of great interest. Innovation prediction has been approached in many ways. There is a set of different mechanisms or technologies to address this question such as the development of models using as input the full text of patents, their metadata or even some indicators associated with them. Hence, our contribution examines patent creation with a new approach based on a combination of scientific articles and patents susceptible to analysis. In order to check the feasibility of this type of study, we have conducted specific research using a precise raw data query on the topic of “Hot stamping” in the Web of Science (WoS). The results were 801 peer-reviewed publications from journals indexed in this database. Moreover, when extending this query, we completed our dataset with an additional 7,610 patents obtained from PatBase®. Our proposed solution, combining both statistical methods and Machine Learning (ML) (Choi et al., <CitationRef CitationID="CR23">2021</CitationRef>) has resulted in a Bayesian network (BN) and a decision tree model that help us to identify scientific production and intellectual property protection behaviour patterns that are able to detect, with high probability, the researcher’s likelihood of going into patenting.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

A new methodology to predict innovation through patents. The hot stamping technology case

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年7月29日

查看全文
  • Luis Miguel Arias,
  • Garikoitz Artola,
  • Javier Nieves,
  • Igone Porto Gomez

摘要

Nowadays, innovation is a key driver of the economy and the competitiveness of organizations, helping them to survive in a global market. The ability to anticipate innovations in the market is thus of great interest. Innovation prediction has been approached in many ways. There is a set of different mechanisms or technologies to address this question such as the development of models using as input the full text of patents, their metadata or even some indicators associated with them. Hence, our contribution examines patent creation with a new approach based on a combination of scientific articles and patents susceptible to analysis. In order to check the feasibility of this type of study, we have conducted specific research using a precise raw data query on the topic of “Hot stamping” in the Web of Science (WoS). The results were 801 peer-reviewed publications from journals indexed in this database. Moreover, when extending this query, we completed our dataset with an additional 7,610 patents obtained from PatBase®. Our proposed solution, combining both statistical methods and Machine Learning (ML) (Choi et al., 2021) has resulted in a Bayesian network (BN) and a decision tree model that help us to identify scientific production and intellectual property protection behaviour patterns that are able to detect, with high probability, the researcher’s likelihood of going into patenting.