Grant: Discovery Grants Program - Individual

Pattern management refers to a set of activities related to the description, manipulation and storage of patterns in a similar (but more elaborated) way as data are managed by database applications. In pattern management and inductive databases, patterns are knowledge artifacts extracted from a data mining process (generally conducted beforehand), and are retrieved upon user request. The objectives of the proposed research program are threefold: (1) handle the problem of pattern manipulation by defining operations on patterns, (2) study the problem of enriching and updating a pattern set (e.g., association rules) when changes occur in the user's needs and the input data (e.g., new objects/attributes, negation or taxonomy), and (3) design procedures and tools for navigating between multidimensional (or even bidimensional) data and related patterns so that data/information can enrich existing patterns and patterns can augment data with knowledge. For the second topic, the problem may be challenging in case of negation, i.e., when the presence as well as the absence of attributes (items) are sought, and when a partial set of patterns is given. Another objective of this research is to carry on our ongoing work on mining and approximating multidimensional data in data warehouses. Since formal concept analysis is an attractive and sound framework for mining concepts and rules, we will use it as the main framework for handling the proposed research activities and we will focus on patterns in the form of association rules and concepts.