A very interesting topic of machine intelligence that has attracted the attention of the research community is multimodal recommendation systems. In recent years, the remarkable growth of social networking sites, the wide-spread use of GPS sensors, and the increasing availability of open geographical databases has motivated a large volume of work on multimodal recommendation systems. This not only opens up a new dimension in for the description, organization, and manipulation of multimedia data, but leverages the deployment of advanced web applications and services in the tourism domain.
Social networking sites offer complementary to content information related to social communities between users, user profiling, metadata related to music, video, images, and text. Geotagging is another information source worth distilling. Personalized recommendation and customized user modelling is crucial for successful social media marketing or link prediction, so novel approaches like hypergraph recommendation, tensor-based recommendation methods, multi-kernel learning, SVM for multi-modal classification etc., that take advantage of the sparsity of multi-link relations/graphs and the context information, can offer promising results.