Abstract
Archives preserve materials that allow us to understand and interpret the past and think about the future. With the evolution of the information society, archives must take advantage of technological innovations and adapt to changes in the kind and volume of the information created. Semantic Web representations are appropriate for structuring archival data and linking them to external sources, allowing versatile access by multiple applications. ArchOnto is a new Linked Data Model based on CIDOC CRM to describe archival objects. ArchOnto combines specific aspects of archiving with the CIDOC CRM standard. In this work, we analyze the ArchOnto representation of a set of archival records from the Portuguese National Archives and compare it to their CIDOC CRM representation. As a result of ArchOnto's representation, we observe an increase in the number of classes used, from 20 in CIDOC CRM to 28 in ArchOnto, and in the number of properties, from 25 in CIDOC CRM to 28 in ArchOnto. This growth stems from the refinement of object types and their relationships, favouring the use of controlled vocabularies. ArchOnto provides higher readability for the information in archival records, keeping it in line with current standards.

Abstract

Abstract
We propose an unsupervised method for propagating automatically extracted fine-grained topic labels among news items to improve their topic description for subsequent text classification procedure. This method compares vector representations of news items and assigns to each news item the label of its closest neighbour with a different topic label. Results obtained show that high precision can be achieved in propagating the top ranked topic label, and that 2-gram and 3-gram feature representations optimize the precision.

Abstract
The study of the blogosphere can provide sociologically relevant data. We analyze the links between blogs in the portuguese blogosphere, in order to understand how they group and interact, to identify clusters and to characterize them. Our data set contains post data for more than 70,000 blogs, with over 400,000 links. The linkage data is represented as a blog graph and partitioned into several slices, according to their in-degree. We then study the evolution of blog features, and observe a consistent pattern of decrease in posting frequency, number of out-links, and post length, as we move from the highly-cited blogs to the less cited ones. Copyright 2010 ACM.

Abstract
We build and analyze a coreference network based on entities from photo descriptions, where nodes represent personalities and edges connect people mentioned in the same photo description. We identify and characterize the communities in this network and propose taking advantage of the context provided by community detection methodologies to improve text illustration and general search. © 2012 Springer-Verlag Berlin Heidelberg.

Abstract