A Keyword Transition Extraction Method for Time-series Text Data and Its Application to Discovering the Transition of Key Technology Elements in Japan
Fan Cheng, Takafumi Nakanishi · 2023
This paper presents a keyword transition extraction method for time-series text data and its application to discovering the transition of key technology elements in Japan. Generally, capturing trends in science and technology is one of the most important issues. To capture these trends, when a method for extracting important keywords year by year from white papers published annually by the government can be established, it becomes possible to visualize the trends in science and technology. This method extracts words from text data for each time, and derives F-Score as a measure of the likelihood of occurrence of each word at that time. By extracting the transition of keywords for each time from the change in F-Score for each word, it is possible to extract the transition of keywords for each time. By realizing this method, it becomes possible to discover important keywords for each era from time-series text data and visualize the transitions of those keywords.