Unification of Higher-level Representative Nodes for Level-Of-Detail of Hierarchical Data Visualization Technique “HeiankyoView”

According to recent growth of information technology, information stored by computers became larger, various, and complicated. Information visualization technique is an active research topic to understand the contents of such information. This thesis discusses the usefulness of JunihitoeView, a technique for visualizing large-scale hierarchical multivariate data. The technique is an extension of HeiankyoView, an existing technique for visualizing large-scale hierarchical data. This thesis shows the results of user tests of JunihitoeView. Based on the discussion of the results of user tests, we found that it may be hard for the users to search the data when the representative icons of the JunihitoeView are unequal. We therefore developed an algorithm to display the equalized representative icons, which show the upper-level of the hierarchy of the data, in HeiankyoView. This thesis presents a technique to display the equalized representative images in CAT, a technique for browsing a large number of images, as an example of the developed algorithm. This technique temporally ignores the numbers of the images in the clusters showed by the representative images, so that it displays all the representative images equally.