Knowledge Discovery in Bibliographic Databases

Citation

de P. Roper, V. (2001), "Knowledge Discovery in Bibliographic Databases", New Library World, Vol. 102 No. 4/5, pp. 180-185. https://doi.org/10.1108/nlw.2001.102.4_5.180.6

The last sentence of the first paragraph of the Introduction, as written by Jian Qin and M. Jay Norton, concludes with “… This issue is devoted to aspects of KDD that are relevant or reflective of the field of library and information”. Then follows a justification for this study of KDD as given by Fayed et al. in 1996. “KDD refers to the overall process of discovering useful knowledge from data and data mining refers to a particular step in this process. KDD encompasses a growing collection of techniques from a variety of disciplines, for investing data to extract knowledge.” There then follows 13 articles of equal obfuscation of the simple principle that very little knowledge is original, but what is original is fresh relationships to existing knowledge. This is a similar principle to the concatenation of knowledge as reflected in citation indexing. A little further in the Introduction is the passage which states “… A similar analogy to Swanson’s can be made in co‐citation passage analysis that, if A is in the starting field, C in the destination field, and B the shared concept/method, then A is to B as C is to B”. I do not feel the world trembling with the originality of a Newtonesque principle, but rather the tedium that comes from the over‐complication of established and accepted ideas. The scholastic impact of these chapters is reduced if you are looking for our world of knowledge and its relationships to be pushed further. But the scholarship of the work cannot be denied. Consolidation of accepted principles, and sometimes the obvious, is always the path of growth.

In the early days of teaching librarianship as an academic study, lecturers were accused of creating their own terminology as a justification for turning techniques into the exclusive and respectable world of academia. It is a pity that this fault still persists in seeking new words to suggest progress, whereas what we have is appreciation of historical data.

The vast majority of the chapters in this issue are given over to the justification of the obvious. For example, Norton, in chapter two, states “… A consideration of KDD database design and cost is data quality. The accuracy of the data’s representation of the entity and environment from which it originated, as well as its currency, are factors of data quality … The lack of, or the failure to apply feedback to data creates a discontinuity between static data and the continually changing world”. The chapter by Kwasnik was perhaps the more readable of all the chapters, but here also the re‐iteration of principles laid down by Ranganathan in the late 1940s and 1950s are already accepted truths, and so the shortcomings of faceted analysis and the resulting classification of materials appreciated. The bibliographic references and their date clearly indicate a grounding in principles but just look at the age of the material cited.

This criticism also applies to the chapter by Qin He on Knowledge discovery through co‐word analysis. There is a good historical basis for the chapter but it takes the study little further than that of an historical study.

The work suffers from too many artificial words and dated bibliographic references, but seen in the light of an historical study of long‐established and accepted working techniques, the work may have literary impact. The economic principle that the total impact of some things being greater than the sum of their parts could be reversed in relation to this publication. Each chapter by itself stands for an example of good academic work, but the proximity of the others detracts from the total work.