This is a massive compendium of survey articles that covers the diverse discipline of data mining and knowledge discovery with a raw patchwork of current materials and timely views.
The editors have succeeded in giving us a snapshot of the current state of the art, including “preprocessing methods, supervised methods, unsupervised methods, soft computing methods, supporting methods, advanced methods, and applications”—all of which relate to specific aspects of a data mining taxonomy, including data mining paradigms.
On one hand, this collection represents millions of man-hours of academic research and industrial development in an area of information technology that is evolving faster than these 120 international specialists can capture in words; thus, much of the nomenclature of the field is author specific, without established standards (this is acceptably consistent with the key journals and major professional conferences in this discipline). On the other hand, the very dynamism of the field leaves the book’s editors with the completely impossible task of organizing the material into a source-book standard reference, a level that they have failed to reach. Essentially, this handbook is really an introduction to the many professional journals and conference papers from which its content was gleaned.
The word “handbook” brings to mind classic texts of engineering, science, and mathematics—most of which are in their umpteenth edition, so this neophyte first edition is a truly noble first approximation of another academic standard. If one were to complement these 1,400 pages of clearly written surveys with basic materials of necessary mathematics, epistemology, linguistics, and the like—and then add a comprehensive handbook-worthy index—the volume might evolve into an encyclopedia. If one were to compress these 1,400 pages by introducing fundamental definitions, which cut across the many subdisciplines—and then redraft all of these articles according to that standard—then one would have advanced the field by a clear generation; this is clearly an impossible pipe dream.
So, having enjoyed the ebb and flow of this first edition, I patiently await the over-the-horizon arrival of the second edition, wondering all the while if the editors will be capable of organizing the myriad of amendments to this massive undertaking, or if some genius of data mining and knowledge discovery will emerge to create cognitive coherence from this jungle of hyperspecialized interdisciplinary information.