2024-2011年上市公司企业渐进式创新数据、渐进式创新锁定数据

Core Insights - The article discusses the measurement of incremental innovation in companies from 2011 to 2024, utilizing patent data to assess the degree of similarity between current and past innovations [1][2]. Methodology - The study employs a method based on machine learning and deep learning for text mining, converting unstructured patent abstracts into numerical vectors to analyze innovation [1]. - A significant focus is on the overlap of content in patent applications, which serves as a key indicator of incremental innovation [1]. - The approach includes the use of TF-IDF weighting to enhance the accuracy of the textual analysis, ensuring that important semantic information is retained [1]. Data Scope - The dataset comprises over 60,000 samples from more than 5,300 companies, providing a comprehensive basis for analysis [1]. - The data collection starts from 2011 due to the availability of patent text data, ensuring a reasonable timeframe for the study [1]. Incremental Innovation Metrics - The article presents a formula to calculate the degree of incremental innovation for companies, based on the similarity of patent texts [2]. - Specific incremental innovation levels for various companies in 2024 are provided, indicating a range of values from 0 to 0.2016 [3].