
Web semantic metadata rules-based harvesting became is an important challenge due to validation of the semantic metadata and the amount of web sites that are rich knowledge sources. Indeed, extracting useful information from the web is the most significant issue of concern for the realization of semantic web; this may be achieved by several ways among which web usage mining, web crawling and scrapping and semantic annotation plays an important role. In this paper, a semantic web metadata harvesting and enrichment model, called Semantic Universal Knowledge Model (SUKM). It goal is to make an enriched semantic encyclopedia. SUKM has to support multi-platform metadata driven applications and interoperability. It may be defined as a structure and rich version of DBpedia in order to increase the usability of various user web knowledge experiences. SUKM aggregates and enriches metadata to create a semantic master metadata catalogue. More specifically, a harvesting modelconsisting of five phases is proposed. This model takes into account sources classification, type of source contents and semantic relationships. SUKM model includes metadata cleaning to remove duplication from different source and semantic metadata enrichments. Semantic Metadata Enrichments consist to identify and enrich topic and emotion metadata hidden within the text or multimedia structure. Enrichment processes use a hybrid machine learning model to propose a topic detection and emotion analysis algorithms. SUKM rules-based harvesting prototype has been implemented using a Java program and more than 10 million metadata hybrid documents have been integrated to the semantic master metadata catalogue.