Want a NOOK? Explore Now
Audience: Data mining academics, research and development professionals in data mining, senior/graduate level students in computer science.
"This book, for the first time, makes it possible to offer Web Mining as a real course." — Professor Jaideep Srivastava, University of Minnesota.
Anonymous
Posted Mon Dec 02 00:00:00 EST 2002
This book describes a range of techniques drawn from academia and experience for discovering resources by crawling the web, and for cluster, classifying and otherwise analyzing these resources. --- Part I is a practical discussion of the infrastructure of search engines, including large-scale web crawlers and information retrieval techniques. Although many existing crawlers originated in universities where they were the subject of academic description, the rise of commercial search engines means that more recent developments are not widely known. Part II discusses a broad range of unsupervised, supervised, and semi-supervised clustering and classification techniques, and Part III explores some applications and outcomes of the ideas developed in the previous sections. In several interesting cases, these applications and outcomes are compared to the (known or suspected) practices of commercial search engines. Having said that, the book emphasizes techniques that do not require that the entire web be mined, only the most relevant parts. --- After the first chapter, most of the material is academic in nature. Topics are described to the level of pseudocode and formulae (with many references to more comprehensive coverage); and the advantages and disadvantages of competing approaches are described, and in many cases demonstrated with practical evaluations. This level of detail is appropriate because the vast size of the web demands that programs be computationally efficient. --- In general, the writing is articulate, the explanations are clear, and the material well-researched. I found the book most valuable for the cohesive overview it provides; its insights into the operation of both large and small-scale crawlers; and its exploration of classification systems that combine the purely textual features of traditional Information Retrieval with link-based features from hypertext graph analysis. I expect that readers with little background in crawling and classification who intend to start Mining the Web will find this an excellent guide; and that experts and academics will find that the author pulls together, discusses and contrasts so many areas of interest, in such detail, as to form a valuable reference.
Was this review helpful? Yes NoThank you for your feedback. Report this reviewThank you, this review has been flagged.
Overview