The Research On Topology Preserving Algorithm For Word Cloud Visualization

As a tool of visualization, word cloud has attracted more and more researchers to study due to its high efficiency and beautiful interface in text analysis. However, it can be imaged that the word cloud may be unstable when changing some words or even all words from it. To solve this problem, this thesis presents two algorithms for two types of word cloud layouts respectively to ensure that the final topology of the cloud will be totally consistent with the original one. After applying these algorithms working, the layout of word cloud will be adjusted dynamically while preserving the user’s metal map during the changing of words in the cloud and this function may be useful when users want to choose one or multiple words to view detail or even delete them. What’s more, the experimental results show that both of the proposed algorithms are effective.The main work of this thesis can be concluded as follows:(1) The current algorithms for word cloud layout are summarized and analyzed. The process of the word cloud pretreatment such as the word extraction and the definition of word’s importance are proposed briefly. Then the existing algorithms are summarized and the advantages and disadvantages of them are analyzed. Furthermore, those algorithms that involve the topology preserving in word clouds are described in detail.(2) The algorithm that preserves the topology in the bounded constraint word cloud is proposed. A word cloud layout method for those words that have certain number with the use of the Power diagram which divides the displace space into certain areas is proposed. Moreover, five kinds of constraint rules, two forces between words and a boundary force are adopted to preserve the topology in the bounded constraint word cloud.(3) The algorithm that preserves the topology in the no boundary word cloud is proposed. The Delaunay Triangulation may be used on the words to obtain the mesh which can make sure the topology not be destructed and then the spring force and repulsion force may be adopted to ensure the layout without overlapping and the words to be compacted. After these operations, the topology of the cloud may be preserved when the words are changed or translated in other languages.