Journal of Digital Information Management 7 6 2009 http://www.dline.info/fpaper/jdim/v7i6/6.pdf This work introduces an approach for identifying duplicated bug reports that combines document indexing and Self-organizing Maps (SOM). The approach consists of a sequence of three processing modules: Preprocessing, Clustering, and Retrieval. The fi rst module produces feature vectors used as description of bug reports. These feature vectors are the input for the second module, clustering, which uses a Selforganizing Map to cluster similar bug reports. Finally, the third module is responsible for fi nding the appropriate cluster for a given new bug report and retrieving the bug reports in that cluster. Experiments with a data base of bug reports demonstrate the usefulness of the proposed retrieval method.