Semantic and Visual Analysis of Metadata to Search and Select Heterogeneous Information Resources

An increasing number of activities in several disciplinary and industrial fields such as the scientific research, the industrial design and the environmental management, rely on the production and employment of informative resources representing objects, information and knowledge. The vast availability of digitalized information resources (documents, images, maps, videos, 3D model) highlights the need for appropriate methods to effectively share and employ all these resources. In particular, tools to search and select information resources produced by third parties are required to successfully achieve our daily work activities. Headway in this direction is made adopting the metadata, a description of the most relevant features characterizing the information resources. However, a plenty of features have to be considered to fully describe the information resources in sophisticated fields as those mentioned. This brings to a complexity of metadata and to a growing need for tools which face with this complexity. The thesis aims at developing methods to analyze metadata easing the search and comparison of information resources. The goal is to select the resources which best fit the user’s needs in specific activities. In particular, the thesis faces with the problem of metadata complexity and supports in the discovery of selection criteria which are unknown to the user. The metadata analysis consists of two approaches: visual and semantic analysis. The visual analysis involves the user as much as possible to let him discover the most suitable selection criteria. The semantic analysis supports in the browsing and selection of information resources taking into account the user’s knowledge which is properly formalized.