Support System for Understanding Unique Information by Visualizing Words that Represent Differences between Text Sets

Abstract

In recent years, the amount of available information has continued to increase. As a result, the utilization of data has become popular, and an environment for extracting valid information from the data is required. Data analysis is based on the comparison of multiple data, but to extract valid information, it is necessary to look through a large amount of data, which requires a lot of time and effort. As data analysis using text, there have been many studies on extracting information from specific texts and on improving the efficiency of information collection by summarizing texts. Unlike those studies, however, there have been few studies on building an environment for comparing two things, which is the basis of data analysis, without limiting the field of data analysis. In this study, we aim to construct a system that can compare two sets of texts on various topics by taking two sets of texts on the matter to be compared as input, visualizing their common and unique words, and displaying the corresponding texts from the text set based on the unique words, thereby supporting the understanding of unique information between the text sets. The objective is to build a system that can compare two sets of texts that can deal with various topics. Through evaluation experiments, we confirmed that the proposed system is effective in comparing two text sets.