Abstract
In this paper we define and analyze stability of a formal concept. A stability index is based on the idea of a dependency in a dataset that can be reconstructed from different parts of the dataset. This idea, underlying various methods of estimating scientific hypotheses, is used here for estimating concept-based hypotheses. Combinatorial properties of stability indices, algorithmic comlpexity of their computation, as well as their dynamics with arrival of new examples are studied.
Users
Please
log in to take part in the discussion (add own reviews or comments).