Supervised cross-domain topic classification trains a model on a labeled source corpus and applies it to an unlabeled target corpus from a different domain. This approach leverages existing labeled data to reduce effort compared with collecting new within-domain training data, while offering clearer, research-targeted topics than unsupervised methods.
Representative democracy bundles many issue positions into party manifestos, and voters ultimately make a single choice between those bundles. That single-choice outcome can hide voters' multidimensional policy preferences. This study uses a conjoint experiment based on real party manifestos from Japan's 2014 House of Representatives election to recover those hidden preferences and to show how voters evaluate whole manifestos.