Word embedding is known as one of the
fundamental tools in natural language processing. Extensive
studies have been done to analyze the performance of word
embedding models in different sentiment analysis tasks.
However, very few works investigated the affective aspects of
these models in emotion analysis area which has been growing
fast recently. This article overviews and analyzes word
embedding specifically for emotion analysis and presents new
research directions accordingly. More particularly, after
investigating the currently affective and non-affective works on
word vectors, a framework for building emotional-aware word
vectors is proposed.