Harvard-McLean Dialogue Analysis

Usage

usage: main.py [-h] [-d] [--liwc]

optional arguments:
  -h, --help   show this help message and exit
  -d, --debug  include debugging messages
  --liwc       export transcripts for LIWC analysis

Environment

This project is written in Python 2.7. An exhaustive list of dependencies can be found in environment.yml.

To create a new conda environment with all requirements, run the following command.

conda env create -f environment.yml

Additionally, install the requirements for deep_disfluency using the following command.

pip install -r deep_disfluency/requirements.txt

To update from an old version of the environment, run the following command.

conda env update environment.yml

Activate/deactivate this environment as usual.

conda activate hm_dialogue
conda deactivate hm_dialogue

Scripts

rename_transcripts.py

Renames new transcript files from Justin to a common format. For example, MULTISENSE_HXXCM_onsiteInterview_audioHeadset_S1+S2_visit3.txt becomes 170208_MS0034.txt.

align_speakers.py

Provides an interface for tagging speaker IDs (e.g., S1) with speaker names (e.g., Participant).

References

Used in the following publications:

A. Vail, E. Liebson, J. Baker, L.-P. Morency. Toward Objective, Multifaceted Characterization of Psychotic Disorders: Lexical, Structural, and Disfluency Markers of Spoken Language. Proceedings of the Twentieth International Conference on Multimodal Interaction (ICMI 2018), Boulder, Colorado, 2018. https://doi.org/10.1145/3242969.3243020

J. Girard*, A. Vail*, E. Liebenthal, K. Brown, C. Kilciksiz, L. Pennant, E. Liebson, D. Öngür, L.-P. Morency, J. Baker. Computational Analysis of Spoken Language in Acute Psychosis and Mania. Schizophrenia Research, 2021. (*equal contribution) https://doi.org/10.1016/j.schres.2021.06.040