Farruque, Nawshad, Osmar Zaiane, and Randy Goebel. "Augmenting semantic representation of depressive language: From forums to microblogs." Joint European conference on machine learning and knowledge discovery in databases. Cham: Springer International Publishing, 2019.

This paper explores the creation of specialized word embedding representations for detecting depressive language in Tweets, particularly when annotated data is limited. The approach starts with general pre-trained embeddings, which are then enhanced with domain-specific embeddings from a smaller dataset. The combined representation better captures the nuances of depression while maintaining broad semantic coverage. Comparative analysis shows that this enhanced word embedding achieves significantly higher performance in detecting depressive Tweets compared to traditional methods such as bag-of-words, sentiment lexicons, psycholinguistic lexicons, and general embeddings.