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

Abstract

Abstract
This paper describes the participation of FEUP, from the University of Porto, in the TREC 2010 Blog Track. FEUP participated in the baseline blog distillation task with work focused on the use of link features available in the TREC Blogs08 collection. The approach presented in this paper uses the link information available in most individual posts to amplify each post's score. Blog scores, and subsequent ranks, are obtained by combining individual post scores. We boost post scores using the in-degree of each post and the h-index of each blog. This results in an improvement of P@10, over our baseline, for the in-degree and the h-index runs. When compared to the in- degree, the h-index run results in higher performance values for each of the applied evaluation metrics.