A Cross-Domain Multimodal Supervised Latent Topic Model for Item Tagging and Cold-Start Recommendation

Abstract

Cross-domain data analysis is playing an increasingly important role in media convergence and can be adopted for many applications. Most existing methods consider the domain discrimination as the multimodal representation difference or the imbalanced item classification distribution, ignoring the different tag semantics among domains. To this end, we propose an explainable cross-domain multimodal supervised latent topic (CDMSLT) model and evaluate our model on two applications. First, we learn a common topic space that is capable of explaining both domain specification and commonality. Second, we apply our model to a multilabel classification task and put forward a cross-domain item tagging method. Third, combining user behaviors and the CDMSLT model, we propose a cross-domain recommendation algorithm that could estimate the user preference on new unseen domains. This article proves the effectiveness of the CDMSLT model by comparing these two applications with existing algorithms in a cross-domain scenario on the Douban dataset.