Abstract:
The thesis is a collection of my original scholarly works published as articles in peer-reviewed scientific journals and books. The articles investigate structural and computational properties for several classes of data dependencies in a variety of data models, including the relational model of data, incomplete data models, complex-value data models, and the eXtensible Markup Language (XML) model. It is also illustrated how the techniques developed in these articles can be applied to reason efficiently about classes of conditional independencies fundamental in artificial intelligence, statistics and probability theory.