The term "semantic network" also refers to co-occurrence networks. In this sense, semantic networks are raphical representations of the co-occurrence of words in units of text (sentences, paragraphs, whole documents).[cite] Each node represents a word in the corpus, and each edge between two nodes represents the frequency (or weighted frequency) with which these two nodes co-occur in the text.

Similar to a vector space model such as Word2vec, the idea is that the relative position of a word in the space reflects its meaning.