Development of iTM: A web-based text mining tool
2016
Park, S.Y. (Tokyo University of Agriculture, Setagaya-ku, Tokyo (Japan)) | Hasebe, T. | Yasue, H.
In recent years, there were attempts of study by using text mining in the agricultural sector in Japan as well. The subjects cover a wide range, but there are still only a few papers. The reasons are that the information system is required for morphological analysis of the text, and digitizing for using extracted morphemes in the statistical analysis is a large-scale work. Therefore, in this paper, we have developed a text mining tool, iTM (internet Text Mining), which can sequentially perform Japanese language morphological analysis, parsing, chi-square test and so on, only by clicking buttons on a webpage that is familiar not only with researchers but also with internet users, such as certified farmers. To confirm the utility of the system, we used a web type questionnaire survey tool that is equipped with iTM, and conducted a case study that is targeting, for free, description answer sentences that have been acquired. As a result, it was revealed that we can feel free to carry out the text mining by using the button click function, which is a feature of iTM, since this function makes it possible to eliminate installations such as a morphological analysis engine and the digitizing works of text data. As an evaluation method for word appearing repeatedly, by applying and comparing with a 1<=Y<2 valuation method which was uniquely devised by us, it has been suggested that it is possible to deepen the consideration for the analysis results.
Показать больше [+] Меньше [-]