Abstract: A data Analysis and visualization technique poses challenges too many areas of the data mining research. There are several visualization techniques and tools have been proposed for almost all domains. But there is a necessity to summarize and visualize a large citation network data according to the user customization. While performing the visualization, influence data should be identified using the summarization technique. The summarization and visualization of graph structured data is a tedious part in research. The existing state of the art influence maximization algorithms can detect the most influential node in a citation network for all structured data, except graph structured data. Clustering techniques are widely used to fold large graph structured data in existing graphs summarization methods. In this paper, first formally define the problem of data summarization and visualization process with three segments, which are effective summarization, localized summarization and handling high, influenced rich information in citation networks. In general, research filed contains lots of interrelated datas, which has multi associations among the data. To handle the above Graph data visualization and summarization problem, here propose a new prototype named as (GSV) Graph Summarization and GSV algorithm for large scale citation networks. Finally present a theoretical analysis on GSV, which is equivalent to the existing kernel k mean clustering algorithm.

Keywords: Data summarization, visual data mining, Graph mining, GSV.