Despite a large number of Knowledge Management (KM) initiatives implemented in organizations, they often fail to manage the natural heterogeneity of organizational knowledge sources. Many approaches to KM have been only based on new information systems technologies to capture all the possible knowledge of an organization into databases that would make it easily accessible to all employees. To overcome heterogeneity, documentation overload and lack of context this article proposes Onto-DOM, a question-answering ontology-based strategy within a Distributed Organizational Memory.

The philosophy regarding knowledge as a "thing" that can be managed like other physical assets has not been quite successful for several reasons related to tacit knowledge capture and tacit-to-explicit knowledge conversion.To overcome some of the most common problems in KM—heterogeneity of knowledge sources, documentation overload and lack of context—we propose Onto-DOM, a question-answering ontology-based strategy implemented within a Distributed Organizational Memory (DOM) that goes beyond information management providing a framework for semantic treatment of organizational knowledge sources.

Onto-DOM is a portable question-answering system that accepts natural language queries and, using a domain ontology, transforms and contextualizes the query eliminating the inherent natural language ambiguity. The system at the same time, recovers those knowledge objects with a higher probability of containing the answer. We say that Onto-DOM is portable because the time needed to implement it in a new domain is minimum, requiring just a change of the associate domain ontology. Onto-DOM combines, in a novel way, a series of techniques to "understand" the natural language query and map it to the semantic annotation done in the organizational knowledge sources.Onto-DOM does an intensive use of domain ontologies in a number of key processes. The domain ontology is used as the core of the classification strategy of knowledge objects. This strategy selects ontological concepts as descriptors obtaining a homogeneous representation of objects structurally heterogeneous. The domain ontology is also used in query refinement, the reasoning process (a process of generalization/specialization using ontology classes and subclasses) and similarity resolution.

Onto-DOM: A Question-Answering Ontology-based Strategy for Heterogeneous Knowledge Sources,ontology, domain, knowledge, objects, natural, capture, heterogeneity, Organizational, overload, documentation, semantic, ambiguity, system, eliminating, heterogeneous, inherent, assets, manage, classification, organizational, physical, probability