Dealing with uncertainties during the data(base) integration process

Applying various data mining techniques to the task of integrating heterogeneous databases into one system leads to the necessity of dealing with uncertain information. The problem of integration is not trivial due to a number of unavoidable conflicts and contradictions. It becomes even more complex, when the initial information is uncertain. We classify three types of uncertainty associated with the integration task and propose a possible procedure of data(base) integration process to deal with the uncertainties.