Why is Metadata a Hot Topic?

A back-of-the-book index and a dictionary are both examples of metadata -- information about information contained in a document or database. Electronic examples of metadata include information encoded in the META tags on Web pages and 'controlled vocabularies,' hierarchical lists of subject terms developed to make commercial bibliographic databases easier to search.

Montague Institute Review (1998). Articles>Knowledge Management>Metadata>Controlled Vocabulary

Identifying Synonymous Concepts in Preparation for Technology Mining    

In this research, the development of a 'concept-clumping algorithm' designed to improve the clustering of technical concepts is demonstrated. The algorithm developed first identifies a list of technically relevant noun phrases from a cleaned extracted list and then applies a rule-based algorithm for identifying synonymous terms based on shared words in each term. An assessment of the algorithm found that the algorithm has an 89-91% precision rate, was successful in moving technically important terms higher in the term frequency list, and improved the technical specificity of term clusters.

Courseault Trumbach, Cherie. Journal of Information Science (2007). Articles>Knowledge Management>Metadata>Controlled Vocabulary

Incremental Maintenance of Generalized Association Rules Under Taxonomy Evolution    

Mining association rules from large databases of business data is an important topic in data mining. In many applications, there are explicit or implicit taxonomies (hierarchies) for items, so it may be useful to find associations at levels of the taxonomy other than the primitive concept level. Previous work on the mining of generalized association rules, however, assumed that the taxonomy of items remained unchanged, disregarding the fact that the taxonomy might be updated as new transactions are added to the database over time. If this happens, effectively updating the generalized association rules to reflect the database change and related taxonomy evolution is a crucial task. In this paper, we examine this problem and propose two novel algorithms, called IDTE and IDTE2, which can incrementally update the generalized association rules when the taxonomy of items evolves as a result of new transactions. Empirical evaluations show that our algorithms can maintain their performance even for large numbers of incremental transactions and high degrees of taxonomy evolution, and are faster than applying contemporary generalized association mining algorithms to the whole updated database.

Tseng, Ming-Cheng, Wen-Yang Lin and Rong Jeng. Journal of Information Science (2008). Articles>Knowledge Management>Metadata>Controlled Vocabulary