What are the recommended approaches for a classification task that involves both structured and unstructured data, with limited dataset size? I’m looking for a generalized model that can be easily customized for various scenarios. Any suggestions on methodologies or specific models that excel in these conditions?