Automatic, Format-independent Generation of Metadata for Documents Based on Semantically Enriched Context Information

摘要: The purpose of this study was to investigate how metadata can be generated automatically for all types documents used in an enterprise, regardless their content. Because the increasing number non-textual documents, i.e. images, audio and video files, full-text indexing is not applicable thus, use has become more important resource description discovery. However, creating manually time consuming error prone moreover barely feasible huge amount enterprise deals with daily. Thus, approach automatic, format-independent metadata generation required. To begin documents’ context analysed. A document considered object, which related other objects such as a task it created for. It recognised that described formally semantically enriched architecture. This architecture then automatic generation. To productive environment determined its linked components, e.g. information stored relational database. Finally procedure setting-up, conducting utilizing generation approach identified. combination these objectives been called mintApproach. With mintApproach system annual economic loss due vast wasted on retrieval addressed. Research design followed deductive mixed method strategy employed, combining four methods: results Representative Study provided comprehensive source analysis creation tools enterprises preferred search strategies. Qualitative interviews conducted survey based a structured questionnaire insights handling enterprise. Action Research prototyping applied two different organisations, non-profit organisation (NPO) domain sexual health small medium-sized (SME), developing contract management software. Evolutionary 'prototyping' built integrated part studies led development executable prototype. Applying enterprises, very business goals, helped avoid common pitfalls like subjectivity, lack of generality replication. The endorsed fact public administrations alike, document’s considered. Although relations between may hidden, low level governance instruments guidelines file storage help reveal relations. For example product or client are implicit structure stored. Determining naming conventions files another way implicitly stating documents. This explicit represented Enterprise Architecture description. found well-known standard Enterprise Architecture modelling, ArchiMate, well suited providing basis core enterprise ontology. ArchiMate refined, enhanced, RDFS-Plus, ontology language machine but also cognitively adequate humans. This enhanced by application specific ontologies reflecting needs, representing knowledge improving lifecycle management. enter rise repository, comprising constituting despite representation. generation context, ontology-to-database-mapping suitable (why used?) The evaluated prototype illustrates scientific models makes easier evaluators assess underlying concepts. Goal evaluation determine appropriateness, capability applicability mintApproach, visualized MeGaWorkbench prototype, assessed appropriate Using promising, particularly regarding multi-media respectively little, meaningless even wrong attributes. beneficial helps meet business needs ever-increasing unstructured reducing personnel involved.