Three methodological results are expected from the presented project. First, a new document hierarchical classification approach that copes with the problem of ID model identification (type, country, year . . . ) with a high number of classes. Second, an adequate ontology for the representation of ID document models. Third, an incremental clustering approach adapted to the problem of forensic link detection from false ID datasets.
- ID Document analysis: Document Extraction and Classification
- ID Document Knowledge Management: Data Modeling
- ID Fraud Pattern extraction: Cluster Analysis and Visualization