Finding the place your data fits with respect to ´ideal knowledge´: a fuzzy process model

It is not easy to define what we mean by knowledge. The term is used loosely or becomes ´defined´ often through context and ability to use ´information´ or an ensemble of lower value elements within a cluster of associated elements to create deeper understanding, and/or new concepts that can lead to high value action. Additionally, semantic identifiers such as data, content, information, and knowledge are often used synonymously, even though there is a sense that there exists a hierarchy of value inherent in the terms. Here we define something we call a data-value chain against which a value scale is mapped in order to conceptualize this value hierarchy and evaluate a data element within it for given input contexts and selection criteria. We further define the position of an element (whether it is a cluster of data or a single atomic datum) in terms of this hierarchy through a value scale, with the added twist that the considered highest apex -knowledge - is not necessarily the most valued in a given context. In order to determine value, multiattribute fuzzy sets and subsets are defined, and it is shown that one can evaluate, define and redefine the position of an element on the data-value chain using one or more value scales. These value scales may be notional or driven by specific input mechanisms that determine context. The approach enables a number of fuzzy operators to be defined that represent value attributes, contexts and selection operations. It is hoped the method will allow rapid classification with the minimum number of attributes, and avoid taxonomy explosions that can plague meta-tagging, for example, as well as efficient classifiers for given contexts.