The transformation method between tree and lattice for file management system

Abstract

To develop an efficient system of file management, tagged image management system based on a lattice structure (TaILaS) has been proposed, and it can avoid the difficulty of classification of files that are able to belong to multiple folders (tree structure). TaILaS assigns huge number of files to concept lattice by formal concept analysis, where the objects and their attributes correspond to files and their tags, respectively. In TaILaS, tagging to files is manually set, and it requires heavy burden of users. In this paper, the translation method from the existing tree structure into the lattice structure is proposed to attach tags automatically and backwards, from the lattice structure into a tree structure. The proposed method uses the names of each folder in the target tree structure as attributes, i.e., the tags of each file of TaILas can be defined as the each folder name. In order to confirm the usefulness of the proposed method, we have performed experiments of translating the tree structure into the lattice structure on a computer (CPU = 2.13 GHz, MM = 4 GB) using the data from Yahoo! Directory and questionnaires. According to the results of the calculation time, the system can work in real-time for 3,000 directories. A subjective questionnaire based analysis concludes that the categorizing images by directories into a lattice structure make file management easy.