Browse Books

Reviewer: Dimitrios Katsaros

The proliferation of computer and communications technologies over the last 20 years has enabled us to produce and store data at an unprecedented pace. The downside of this revolution is that we now need sophisticated techniques (algorithms) to extract useful information from the raw data. Various data mining techniques have been developed for this goal, that is, to discover valid, novel, and potentially useful patterns in the data. These techniques involve the discovery of patterns, associations, changes, outliers, and statistically significant complex structures in huge collections of data. There are several success stories of data mining research and development reported in the literature, such as fraud detection and the discovery of frequent itemsets-both sequential and nonsequential. Frequent itemset discovery algorithms can extract co-occurrences of items that can take into account the ordering of items or not. Even though the use of sets (or sequences) enabled the modeling of many application domains, such as market basket analysis, several applications have emerged whose data models do not fit the concept of a set or sequence, but require the deployment of richer abstractions, such as graphs. Graphs arise in a number of different domains, including network intrusion detection, Web 2.0, very-large-scale integration (VLSI) reverse engineering, social network analysis, search engine ranking, chemical compound and protein classification, and many more. Thus, the need to extract complex graph-like patterns in massive data collections has become a necessity. This book provides a survey of some recent advances in graph mining. It contains chapters on graph languages, indexing, clustering, pattern mining, keyword search, and pattern matching. Additionally, it studies aspects of graph mining in areas including stream mining, Web graphs, social networks, and biological data. The chapters are written by prominent researchers who have done extensive work in their respective fields. Reading this book along with Mining graph data [1] offers a solid background to graph data management and mining. The book is comprised of 19 chapters. With the exception of the introductory one, each chapter is a comprehensive survey of some graph data analytics. Chapter 2 surveys the algorithms and applications of graph data mining in general. Chapter 3 presents the power laws that govern some properties of the graphs, such as diameter and node degrees, and describes graph generators that can be used to produce graphs with specific characteristics. Other chapters take a more database-oriented approach, studying problems related to querying and indexing graph-based data. Specifically, chapter 4 presents the GraphQL query language and its evaluation, chapter 5 focuses on feature-based graph indexing, and chapter 6 studies the processing of reachability queries. Chapter 7 deals with exact and approximate graph matching, and chapter 8 discusses keyword search in Extensible Markup Language (XML) graph data. Chapters 9 to 12 survey algorithms for graph clustering, dense subgraph discovery, graph classification, and frequent subgraph mining, respectively. The rest of the chapters deal with modern topics: chapter 13 studies streaming graphs, such as those that arise in search engines and telecommunications call graphs; chapter 14 studies privacy preservation; and chapters 15 and 16 focus on graph mining for the Web and social networks. Finally, chapters 17 to 19 examine graph problems in the areas of software engineering and biological and chemical databases. The book is targeted at advanced undergraduate or graduate students, faculty members, and researchers from both industry and academia. The necessary background of these individuals should cover basic data mining concepts and graph theory. As expected, the reader will not obtain a thorough understanding of the entire field-the field of graph mining is extremely dynamic, as new knowledge is added every year from the many papers published at conferences and in journals. Thus, the book aims to provide breadth, rather than become a complete reference. Overall, I highly recommend this book to someone who is starting to explore the field of graph mining or wants to delve deeper into this exciting field. Online Computing Reviews Service