This repository introduces the persuasiveness metric.
- The persuasiveness metric is concluded from a systematic literature review on persuasive natural language processing. It quantitatively determines the persuasiveness of any given email.
- This persuasiveness metric scores a message through a comparison to 20.000 successful messages. These were extracted, scored and compiled from the (GMANE) email dataset. It consists of 160 million emails.
- Additionally, this persuasvieness-metric provides a word level analysis of the text, and gives descriptive indices of the input email. These indices are inspired by the Text Complexity Analyzer implementation by Hans.
- This object oriented project is based on SpaCy v3, and can be deployed as a FastAPI microservice on Google Cloud Run.
Please contact me in case you need any information or have any questions.
Instructions to run the persuasiveness metric locally:
git clone https://github.com/seduerr91/persuasiveness-metric
cd persuasiveness-metric
pip3 install -r requirements.txt
python3 app.py