Publication at Faculty of Mathematics and Physics |
2014
Abstract
In this paper, we introduce the Visual History Archive of the USC Shoah Foundation as a multimodal data resource for sentiment analysis in Czech, but potentially all thirty three languages it contains. We take the opportunity of having both physical access to these unique data and the well-established research group on sentiment analysis at Charles University in Prague.
Our aim is to provide methodology for sentiment annotation of these multimodal data combining subjectivity detection within a treebank with spoken term detection.