A Web Site Misinformation Detection System

"The proposed project concerns the design and development of a Web Site MisinformationDetection System for automatically reviewing web sites to identify intentionallymisleading information. In the proposed approach, a relational ontology of website categories and semantic entities is populated by autonomous agents capable ofsearching the web for documents of preselected domains and extracting semanticinformation from their content. Supervised learning techniques are employed toclassify and differentiate web sites and document entities by their statistical andrelational charecteristics. The system's acquired knowledge of domain-specific website characteristics and patterns is then used to identify anomalies andinconcistencies in novel web sites. Web sites exhibiting a high degree of deviationfrom characteristics more typical of their category are further subjected to temporalmonitoring, to identify temporal inconsistencies and confirm the presence offraudulent information. Fraud detection; Supply chain; Supplier Qualifications"