Network semantici e recupero di database obsoleti:
Research
Network semantici e recupero di database obsoleti:
il caso di Patrimonio SOS
The project, strongly interdisciplinary in nature, is aimed at defining analytical criteria and methods for the retrieval of data belonging to obsolete databases. In addition to being part of the activities of the Interdepartmental Strategic Plan of the Department of Humanities and Cultural Heritage (DIUM), the research also falls within the interests of the Interdepartmental Center Artificial Intelligence for Cultural Heritage (AI4CH) of the University of Udine, whose goal is the enhancement of cultural heritage through the development of new conceptual models and new interdisciplinary tools.
Objectives
The project set out to investigate the possibility of recovering databases that are no longer accessible due to software obsolescence or structural issues. This is a highly topical issue, which in recent years has led to the loss of numerous products of humanities research.
In particular, the research focused on the specific case of the Patrimonio SOS database, which at the start of the project had been migrated to the E-Dvara platform but was neither accessible nor navigable. Comprising more than 170,000 press review records, this database is of exceptional importance and uniqueness, as it provides a systematic perspective on issues related to the protection of cultural heritage over a period of more than two decades (2002–2020).
Results in the field of DH
Not only did the project enable the recovery and public availability of the data, but it also enhanced their value by creating new and more fine-grained research functionalities. During the course of the research, data mapping and analysis procedures were investigated both at the semantic level and at the formal and descriptive level: in the former case, in order to contextualize the value of the data and reconstruct the structure of the original database, thereby optimizing its usability; in the latter, to enable the correct normalization of attributes and XML elements. In addition, data mining processes and machine learning algorithms were applied for the recognition and identification of unstructured data.
Confirming its interdisciplinary nature, the project involved close collaboration with Federico Giubbolini (within the framework of a research fellowship supervised by Donata Levi) and with the student Laura Pagotto, who addressed specific aspects of the research in her computer science degree thesis, supervised by Andrea Brunello.
