Inducing Domain-Specific Sentiment Lexicons from Unlabeled Corpora

Understand

A word's sentiment depends on the domain in which it is used.

• Computational social science research thus requires sentiment lexicons that are specific to the domains being studied.
• We combine domain-specific word embeddings with a label propagation framework to induce accurate domain-specific sentiment lexicons using small sets of seed words, achieving state-of-the-art performance competitive with approaches that rely on hand-curated resources.
• Using our framework we perform two large-scale empirical studies to quantify the extent to which sentiment varies across time and between communities.