andreasvc/litvecspace

Accompanying code for the paper "Vector space explorations of literary language"

Vector space explorations of literary language

Accompanying code for the paper "Vector space explorations of literary language"

This repository is intended for documentation purposes, as the relevant data cannot be made publicly available.

Data used

• The Riddle of Literary Quality corpus of 401 novels and survey data; http://literaryquality.huygens.knaw.nl/
• Topic model presented in http://dh2016.adho.org/abstracts/95

Requirements

See requirements.txt. Install with pip3 install -r requirements.txt

Overview

• features.py: divides corpus into chunks and computes BoW features and paragraph vectors
• notebook.ipynb: fit predictive models, data analysis, tables, figures, etc.

Reference

@article{vancranenburgh2019vecspace,
    author={van Cranenburgh, Andreas
            and van Dalen-Oskam, Karina
            and van Zundert, Joris},
    title={Vector space explorations of literary language},
    year={2019},
    journal={Language Resources and Evaluation},
    month={Feb},
    day={09},
    issn={1574-0218},
    doi={10.1007/s10579-018-09442-4},
    url={https://doi.org/10.1007/s10579-018-09442-4}
}