Related Experiment Videos
The NITE XML Toolkit: flexible annotation for multimodal language data
Jean Carletta1, Stefan Evert, Ulrich Heid
1Human Communication Research Centre and Language Technology Group, University of Edinburgh, Edinburgh, Scotland. j.carletta@edinburgh.ac.uk
Behavior Research Methods, Instruments, & Computers : a Journal of the Psychonomic Society, Inc
|November 1, 2003
Summary
Researchers need flexible coding tools for multimodal corpora. The NITE XML Toolkit provides library support for creating structured, time-stamped codes, enabling richer analyses of human communication.
Area of Science:
- Computational linguistics
- Human-computer interaction
- Corpus linguistics
Background:
- Multimodal corpora of human interaction are increasingly available.
- Current analysis tools often focus on either timing/sequencing or linguistic structure, but not both.
- A need exists for integrated tools that capture both temporal and structural aspects of communication.
Purpose of the Study:
- To introduce the NITE XML Toolkit as a solution for analyzing multimodal corpora.
- To enable the development of flexible, research-driven coding schemes for linguistic data.
- To facilitate richer and more comprehensive analyses of human communication.
Main Methods:
- The NITE XML Toolkit provides library support for software developers.
- It includes data handling, a query language with temporal and structural constructs, and components for graphical interfaces.
- Sample programs and an experimental engine for building interfaces are also provided.
Main Results:
- The toolkit enables the creation of code sets with both timings and structure.
- It supports flexible design to address specific research questions.
- Facilitates richer analyses than previously possible.
Conclusions:
- The NITE XML Toolkit empowers researchers to conduct more in-depth analyses of multimodal communication.
- It bridges the gap between temporal and structural analysis of linguistic data.
- Promotes the development of advanced tools for understanding human interaction.