Globally, the rapid increase in patents, recognized as documents with high potential value, highlights their significance. These patents possess characteristics of big data, enabling the creation of diverse values through analysis.
While statistical...
Globally, the rapid increase in patents, recognized as documents with high potential value, highlights their significance. These patents possess characteristics of big data, enabling the creation of diverse values through analysis.
While statistical techniques and machine learning are crucial for patent analysis, gaining insights from the results is equally important. Currently, the visualization of patent data predominantly involves simplistic forms. This study proposes circular packing as a visualization method that utilizes topic modeling to quantitatively identify the technologies in patent big data and understand technological development trends. Using LDA topic modeling and circular packing, we analyzed 5,172 patents related to mild hybrid technologies filed with the United States Patent and Trademark Office from 2005 to 2019.
The results identified major applicants like Toyota Group, Hyundai Group, Ford, GM, and Honda as key filers in the mild hybrid automobile sector. The mild hybrid automobile technology comprises 11 technology groups, including predictive modeling and signal processing, thermal and fluid systems, and genetic and chemical analysis. Among these, predictive modeling and signal processing, and thermal and fluid systems were identified as the key technologies, likely reflecting their role in enhancing automotive efficiency and proactive compliance with environmental regulations.
Moreover, while Hyundai emerged as a major filer, the lack of technically robust patents suggests a need for the company to file patents in key technology groups to enhance competitiveness.
This research proposes a methodology for analyzing and visualizing the trends in mild hybrid automobile technologies using patent data, enabling intuitive understanding of patent trends.