Research HighlightsExtracting "Information" From TextKazuaki KishidaDepartment of Library and Information ScienceDec. 21, 2020
My research is aimed at making improvements to information retrieval techniques. While information is now widely available via online search engines, before we had to visit the library in order to search for information. This has a historical precedent in the information retrieval systems created when the pre-computer search mechanisms used by libraries were first converted for digital use. In this respect, information search is at the core of library and information science, and the improvement of its technology is an important research topic.
Information retrieval research underwent a drastic shift in the early 1990s. Effective and efficient search of internet information resources requires advanced information retrieval technology, which led experts in database management and natural language processing to work together to promote international research projects. Until then, the object of information retrieval in the field of library and information science was not the information itself but library materials such as books and journals. In other words, until then, we were searching for the container of information, not for the information itself, something that has changed dramatically since the 1990s.
Today, with the development of computer technologies, we are able to search inside of these materials. In other words, if a document contains a series of sentences, we are now able to reach in and analyze that text data directly. As a result, attention has been given to “text mining,” the process of "digging out" information from within a text. This process relies heavily on information retrieval and natural language processing. To that end, I am working on the issue of automatically classifying documents and words into similar data sets or clusters. For example, we can learn about topics trending around the world through the automatic classification of the large number of tweets that are posted each day. And it is now possible to narrow down and extract opinions and thoughts about a specific topic. I believe that, with further improvements, this technology has great potential for a large number of other beneficial applications for text data analysis.
(2020/12/21)
