On Association, Similarity and Dependency of Attributes

Abstract

Association, similarity, and dependency of attributes represent useful information and knowledge that can be derived from data sets. Similarities indicate the closeness of attributes reflected by their values on a set of objects. Two attributes are similar if every object is likely to have the same value on them. Data and functional dependencies show the connection and association between attributes. They are characterized by the problem of determining the values of one set of attributes based on the values of another set. Two levels of dependencies, referred to as the local and global dependencies, may be observed. The local dependencies show how one specific combination of values on one set of attributes determines one specific combination of values on another set. The well known association rules, which state the presence of one set of items implies the presence of another set of items, may be considered as a special kind of local dependencies. The global dependencies show all combinations of values on one set of attributes determine all combinations of values on another set of attributes. Functional dependencies in relational databases are examples of global dependencies.