An Approach for the Extraction of Information from Heterogeneous Sources of Textual Data