Journal "Software Engineering"
a journal on theoretical and applied science and technology
ISSN 2220-3397
Journal "Software Engineering"
a journal on theoretical and applied science and technology
ISSN 2220-3397
Issue N9 2015 year
In the course of daily activities, one is confronted with increasingly volume of information, a large part of which is stored in digital format. In this paper the MemoPIM system is presented. MemoPIM supports management of personal information and personal knowledge and develops ideas of the Semantic Desktop <97> an approach to the personal information space organization in accordance with principles of Semantic Web and Linked Open Data.
Knowledge and information is stored in RDF repository, based on OWL ontology. Ontology describes classes from different domains, such as file system ontology, publication ontology, calendar ontology etc. Knowledge can be imported from external sources or created manually by user. System provides adapters for following sources: file system, email, calendar, browser bookmarks. Imported information is transformed to RDF in accordance with system ontology. In some cases additional metadata is created, for example author, annotation and title is extracted from scientific articles.
During user work, actions performed by him are logged to knowledge base. These logs are later used for several tasks such as context inference, knowledge base analysis and tag recommendation. Context inference is based on spreading activation algorithm over RDF graph. Activation values for RDF nodes are recalculated after each user action. To create additional relations in RDF data, knowledge base analysis is performed. Another service provided by system is automatic categorization of resources. Categorization is based on tag recommendation algorithm.
System provides two types of user interfaces for knowledge management <97> based on hierarchical presentation of data and based on mind maps.