Sparse computation for large-scale binary classification

Well-known data mining algorithms rely on inputs in the form of pairwise similarities between objects. For large datasets it is computationally impossible to perform all pairwise comparisons. We therefore propose a novel approach that uses approximate Principal Component Analysis to efficiently identify groups of similar objects. The effectiveness of the approach is demonstrated in the context of binary classification using the supervised normalized cut as a classifier. For large datasets from the UCI repository, the approach significantly improves run times with minimal loss in accuracy.

SLR-GNSS analysis in the framework of the ITRF2013 computation

Pre-combined SLR-GNSS solutions are studied and the impact of different types of datum definition on the estimated parameters is assessed. It is found that the origin is realized best by using only the SLR core network for defining the geodetic datum and the inclusion of the GNSS core sites degrades the origin. The orientation, however, requires a dense and continuous network, thus, the inclusion of the GNSS core network is absolutely needed.

“Facebook Style”: The Use of Emoticons and Graphic Signs by Users in the Cape Flats

Over the last few years Facebook has become a widespread and continuously expanding medium of communication in Africa and worldwide. Being a new medium of social interaction, Facebook produces its own communication style. It is a style conditioned by the medium and the community of users. My focus of analysis is how Facebook users from the city of Cape Town create this style by means of emoticons and other graphic signs in order to reflect the reality of living in Cape Town’s underprivileged areas. This study is based on a theoretical framework which combines sociolinguistics with Computer-Mediated-Communication to study the emergence of a style peculiar of the online social networks. In a corpus of Coloured Facebook users from the Cape Flats, I have analysed the emergence of emoticons and other graphic signs related to Capetonian gang culture and then tracked the spread of these features to the extensive use by users not related to gangs. It can be deduced that in this process the analysed features amplify their meaning and are employed in a much broader context as their original use. Due to the development and spread of these features we can consider the peculiar electronic communication of Facebook as a style constrained by the electronic medium and its users. It is a style which serves the users to create social meaning and to express their linguistic identities.