Besides finding changing areas based on the number of node and edge evolutions, a more interesting problem is to analyze the impact of these evolutions to graphs and find the regions that exhibit significant changes when these evolutions happen. The more different the relationship between nodes in a certain region is, the more significant this region is. This problem is challenging since it is hard to define the range of changing regions that is closely related to actual evolutions. We formalize the problem by using a similarity measure based on neighborhood random walks, and design an efficient algorithm which is able to identify the significant changing regions without recomputing all similarities. Meaningful examples in experiments demonstrate the effectiveness of our algorithms. / Graph patterns are able to represent the complex structural relations among objects in many applications in various domains. Managing and mining graph data, on which we study in this thesis, are no doubt among the most important tasks. We focus on two challenging problems, namely, graph summarization and graph change detection. / In the area of summarizing a collection of graphs, we study the problem of summarizing frequent subgraphs, since it is not much necessary to summarize a collection of random graphs. The bottleneck for exploring and understanding frequent subgraphs is that they are numerous. A summary can be a solution to this issue, so the goal of frequent subgraph summarization is to minimize the restoration errors of the structure and the frequency information. The unique challenge in frequent subgraph summarization comes from the fact that a subgraph can have multiple embeddings in a summarization template graph. We handle this issue by introducing a partial order between edges to allow accurate structure and frequency estimation based on an independence probabilistic model. The proposed algorithm discovers k summarization templates in a top-down fashion to control the restoration error of frequencies within sigma. There is no restoration error of structures. Experiments on both real and synthetic graph datasets show that our framework can control the frequency restoration error within 10% by a compact summarization model. / The objective of graph change detection is to discover the changing areas on graphs when they evolves at a high speed. The most changing areas are those areas having the highest number of evolutions (additions/deletions) of nodes and edges, which is called burst areas. We study on finding the most burst areas in a stream of fast graph evolutions. We propose to use Haar wavelet tree to monitor the upper bound of the number of evolutions. Our approach monitors all potential changing areas of different sizes and computes incrementally the number of evolutions in those areas. The top-k burst areas are returned as soon as they are detected. Our solution is capable of handling a large amount of evolutions in a short time, which is consistent to the experimental results. / The objective of graph summarization is to obtain a concise representation of a single large graph or a collection of graphs, which is interpretable and suitable for analysis. A good summary can reveal the hidden relationships between nodes in a graph. The key issue of summarizing a single graph is how to construct a high-quality and representative summary, which is in the form of a super-graph. We propose an entropy-based unified model for measuring the homogeneity of the super-graph. The best summary in terms of homogeneity could be too large to explore. By using the unified model, we relax three summarization criteria to obtain an approximate homogeneous summary of appropriate size. We propose both agglomerative and divisive algorithms for approximate summarization, as well as pruning techniques and heuristics for both algorithms to save computation cost. Experimental results confirm that our approaches can efficiently generate high-quality summaries. / Liu, Zheng. / Advisers: Wai Lam; Jeffrey Xu Yu. / Source: Dissertation Abstracts International, Volume: 73-06, Section: B, page: . / Thesis (Ph.D.)--Chinese University of Hong Kong, 2011. / Includes bibliographical references (leaves 133-141). / Electronic reproduction. Hong Kong : Chinese University of Hong Kong, [2012] System requirements: Adobe Acrobat Reader. Available via World Wide Web. / Electronic reproduction. [Ann Arbor, MI] : ProQuest Information and Learning, [201-] System requirements: Adobe Acrobat Reader. Available via World Wide Web. / Abstract also in Chinese.
Identifer | oai:union.ndltd.org:cuhk.edu.hk/oai:cuhk-dr:cuhk_344974 |
Date | January 2011 |
Contributors | Liu, Zheng, Chinese University of Hong Kong Graduate School. Division of Systems Engineering and Engineering Management. |
Source Sets | The Chinese University of Hong Kong |
Language | English, Chinese |
Detected Language | English |
Type | Text, theses |
Format | electronic resource, microform, microfiche, 1 online resource (xiv, 141 leaves : ill.) |
Rights | Use of this resource is governed by the terms and conditions of the Creative Commons “Attribution-NonCommercial-NoDerivatives 4.0 International” License (http://creativecommons.org/licenses/by-nc-nd/4.0/) |
Page generated in 0.0013 seconds