Dialogue Act Classification
Also known as: dialogue act tagging, speech act classification, Diyalog Eylem Sınıflandırma (Dialogue Act Classification)
Dialogue act classification is a natural-language-processing task that automatically labels the communicative function of each utterance in a conversation — such as question, answer, greeting, or rejection. Consolidated by Jurafsky et al. (1997) and Stolcke et al. (2000), it is a foundational component for chatbots and discourse analysis.
Read the full method
Sign in with a free account to read this section.
Method map
The neighbourhood of related methods — select a node to explore.
When to use it
Use dialogue act classification when you have conversational or text-based dialogue data and need to label what each utterance does. It requires dialogue data to be available and a dialogue act taxonomy (such as DAMSL or ISO 24617-2) to be chosen up front. A modest corpus — on the order of a few dozen utterances or more — is needed; with no dialogue data the method cannot run.
Strengths & limitations
- Turns free-form conversation into structured, per-utterance functional labels that downstream systems can reason over.
- Builds on standardised taxonomies (DAMSL, ISO 24617-2), so labels are comparable across studies and systems.
- Foundational for conversational interfaces and discourse analysis, fitting naturally into chatbot and dialogue-system pipelines.
- Depends on dialogue data being available and on a taxonomy being fixed in advance.
- The function of an utterance often depends on conversational context, which models that ignore context handle poorly.
- Rare or ambiguous acts are easily confused with more frequent ones, so a single overall accuracy figure can hide weak spots.
Frequently asked
What is a dialogue act?
A dialogue act is the communicative function of an utterance — what the speaker is doing with that turn, such as asking a question, giving an answer, greeting, or rejecting. Dialogue act classification assigns one such label to each utterance from a fixed taxonomy.
Which taxonomy should I use?
Choose an established scheme before labelling, such as DAMSL or the ISO 24617-2 standard. The classifier can only assign acts from the inventory you define, so fixing a consistent taxonomy up front is essential.
Why does conversational context matter?
The same words can perform different acts depending on what preceded them — for example, an utterance may be a question in one context and an answer in another. Models that account for dialogue history generally tag more accurately than ones that judge each utterance in isolation.
How is performance evaluated?
Predicted acts are compared against labelled dialogue data with accuracy and per-class metrics, and confusions between similar acts are inspected. On imbalanced act distributions, overall accuracy alone can hide systematic errors on rarer acts.
Sources
- Stolcke, A. et al. (2000). Dialogue Act Modeling for Automatic Tagging and Recognition of Conversational Speech. Computational Linguistics, 26(3), 339-373. DOI: 10.1162/089120100561737 ↗
- Jurafsky, D. et al. (1997). Automatic Detection of Discourse Structure for Speech Recognition and Understanding. ICASSP. DOI: 10.1109/asru.1997.658992 ↗
How to cite this page
ScholarGate. (2026, June 1). Dialogue Act Classification. ScholarGate. https://scholargate.app/en/text-mining/dialogue-act-classification
Which method?
Set this method beside its closest kin and read them side by side — the library lays the books on the table; the choice is yours.
- Intent DetectionText mining↔ compare
- Sentiment AnalysisText mining↔ compare
- Slot FillingText mining↔ compare
- Text ClassificationText mining↔ compare