Graph Interpretation, Summarization and Visualization Techniques: A Review and Open Research Issues

Graphs has been a ubiquitous way of representing heterogeneous data. There are many studies focused on graph learning highlighting the approaches for graph data extraction, interpretation and graph summarization. Graph data summarization is achieving more expansion due to the broader length of sizeable applications and interpretation of proper understanding about the hidden details of the data using deep learning-based graph representation. Graph interpretation and summarization have come up as an interdisciplinary room that has vividly broader influence over multiple parallel areas and real-world applications. In other words, extraction of relevant data from massive and complex graph structure, enables the data to be used by many application area. However, it is found that recognizing the discriminatory and hidden properties from massive heterogeneous data is not easy in case of both nodal graph and graph image (also known as chart image). Hence, deep learning based approaches eventuated as a satisfactory solution. This paper presents an outline of the quantitative and statistical approaches used for learning and understanding different integrant of nodal graph and information graph, such as data extraction and processing, interpretation, summarization and visualization, by using graph-based learning methods. These integrant are broadly considered under (or as) SIV Model in this paper. Paper also discusses the influence of summarization techniques on the visualization of large data graphs and upcoming research areas of summarization. Lastly, paper provides with brief overview of challenges, application area, benefits of graph interpretation, summarization, and visualization, while providing existing tools and datasets available for graph processing and learning.