https://github.com/rmitsch/righteous-mind
Approximative empirical verification of the moral-based framework to distinguish between different political preferences as introduced in "The Righteous Mind" by Jonathan Haidt (2012).
https://github.com/rmitsch/righteous-mind
bert morality natural-language-processing python twitter word-embeddings
Last synced: about 1 month ago
JSON representation
Approximative empirical verification of the moral-based framework to distinguish between different political preferences as introduced in "The Righteous Mind" by Jonathan Haidt (2012).
- Host: GitHub
- URL: https://github.com/rmitsch/righteous-mind
- Owner: rmitsch
- License: gpl-3.0
- Created: 2018-12-13T13:51:10.000Z (over 7 years ago)
- Default Branch: master
- Last Pushed: 2021-05-18T09:56:42.000Z (about 5 years ago)
- Last Synced: 2025-06-13T22:43:11.349Z (12 months ago)
- Topics: bert, morality, natural-language-processing, python, twitter, word-embeddings
- Language: Jupyter Notebook
- Homepage:
- Size: 8.91 MB
- Stars: 2
- Watchers: 2
- Forks: 1
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
### Politics, Morals and Tweets: Identifying Political Affiliation Utilizing Moral Foundations Theory and Contextual Embeddings
This project seeks to extract moral preferences from US politicians' tweets and predict their political ideology based on these moral preferences. The underlying theoretical framework is the _[Moral Foundations Theory](https://en.wikipedia.org/wiki/Moral_foundations_theory)_.
------
#### Notes
* Dependencies can be found in `environment.yml`.
* Starting BERT server for inferring seed phrases for moral values (i. e. we want one embedding vector for the whole phrase):
```bert-serving-start -model_dir ~/Development/data/BERT/base/ -num_worker=1 -cpu```
* Starting BERT server for inferring words in tweets (i. e. we want one embedding vector per token):
```bert-serving-start -model_dir ~/Development/data/BERT/base/ -num_worker=1 -cpu -max_seq_len=40 -pooling_strategy NONE```
* Created for the course "Social Media Data: Quantitative Text Analysis of Big Data", University of Vienna, Winter Semester 2018/2019.