Automatic Evaluation of Local Topic Quality
This work addresses the need for better evaluation of topic models for researchers and practitioners in natural language processing, though it is incremental as it builds on existing topic modeling frameworks.
The paper tackles the problem of evaluating token-level topic assignments in topic models, which are important for downstream tasks but previously only assessed with global metrics. They propose a new automated metric called consistency, based on topic switches, which correlates most strongly with human judgments of local topic quality.
Topic models are typically evaluated with respect to the global topic distributions that they generate, using metrics such as coherence, but without regard to local (token-level) topic assignments. Token-level assignments are important for downstream tasks such as classification. Even recent models, which aim to improve the quality of these token-level topic assignments, have been evaluated only with respect to global metrics. We propose a task designed to elicit human judgments of token-level topic assignments. We use a variety of topic model types and parameters and discover that global metrics agree poorly with human assignments. Since human evaluation is expensive we propose a variety of automated metrics to evaluate topic models at a local level. Finally, we correlate our proposed metrics with human judgments from the task on several datasets. We show that an evaluation based on the percent of topic switches correlates most strongly with human judgment of local topic quality. We suggest that this new metric, which we call consistency, be adopted alongside global metrics such as topic coherence when evaluating new topic models.