Search

This paper presents an architecture for semantic context-aware recommender systems intended to be implemented on organizational settings. We described the use of this architecture in the development of the Find-It! project to validate its utility. The Find-It! project consider contextual factors to recommend people, knowledge objects, places, activities, events, technological resources and services to the visitors of an organization. The architecture defines an inference module to filter the items subject of recommendation by using different contextual dimensions that are modeled in a single ontologies, the novelty of an item, users' perception of recommendations and contextual conditions considered when they were recommended in previous visits. We also defined a mobile client interface capable to adapt to the user characteristics and device. This client interface counts with implicit and explicit mechanisms to identify the user satisfaction degree with the recommendations received and its quality. Experience obtained with the development of the Find-It! project shows a satisfactory performance. In this way, it is adequately considered to use the architecture to develop a wide range of organizational recommender systems. [ABSTRACT FROM AUTHOR]