A Framework for the Construction of Monolingual and Cross-lingual Word Similarity Datasets