Mateusz Dziedzic

The aim of the project was to incorporate bipolarity of user intentions and evaluations into flexible database querying and, first of all, linguistic data summarization which provide for a flexible data mining techniques easily comprehensible even by an untrained end-user.

One of the main tracks of research followed the idea of bipolar database queries which allow for a separate formulation of the negative (mandatory) and the positive (optional) conditions of a non-complementary character. Two operators, the "and possibly" (initially discussed by Lacroix and Lavency for a crisp case in 1987) and "or if impossible" (related to the “or else” operator proposed by Lietard et al. in 2011), were proposed, which model the aggregation of bipolar conditions taking into account a possibility or impossibility (with respect to the data present in a data collection under consideration) of satisfying one them. Those proposals were further developed into two types of contextual bipolar queries, where the possibility or impossibility of fulfilling one of the conditions were referred to only a part of the data, defined as a context of the tuple being under consideration (e.g., one may want to consider employees from every department separately to decide if a given condition is possible or not to be satisfied).

Above results concerning bipolarity in queries have been adopted to develop a new concept of the linguistic summaries of data. A specific relation between the flexible database querying and linguistic summarization was employed and an idea of contextual linguistic summary was proposed and studied, e.g., in the scope of its quality criteria.

The above mentioned research has been carried out in the close cooperation with DDCM group from Ghent University (Belgium). This cooperation resulted also in a study on a bipolar approach to summarizing incomplete data. All results are documented in a number of papers, a list of which may be find here.

You are here: Home Results Mateusz Dziedzic