Skip to content

Analyze autonomous vehicles (AVs) user sentiment from online discussion forums to predict commercial adoption using news.

Notifications You must be signed in to change notification settings

ZachariyaSow/dynamic-sentiment-AVs

Repository files navigation

dynamic-sentiment-AVs

🎯 Description

An end-user sentiment assessment tool to predict autonomous vehicles (AVs) adoption and commercialization.

💭 Scope definition and purpose

Our study complements previous research on social media platform user sentiment on AVs by (a) introducing a dynamic component by observing sentiment across time and critical news events, (b) correlating key user-discussed topics with shifts in user sentiment, and (c) predicting user sentiment based on news’ underlying topic assessment. We will leverage NewsCatcherAPI to extract filtered topic-specific news posts from the last 5 years to train a model using topic scores for sentiment prediction.

🎓 Context

Our research is part of a project in a text mining class at McGill University.

💻 Technology and techniques

  • Python
  • R
  • Sentiment analysis using VADER and TextAnalyzer
  • GPT-augmented topic modeling
  • Regression inference (ridge)
  • Deep learning models (RNN)
  • Reddit API
  • NewsCatcherAPI for news scraping to extract topic-specific media postings and releases for model training data.

✍🏽 Author(s)

A project by Zachariya Sow, Paul Prindiville-Porto, Gabriel Abbas, and Artiom Bakhrakh.

✉️ Contact information

📖 Previous literature

About

Analyze autonomous vehicles (AVs) user sentiment from online discussion forums to predict commercial adoption using news.

Topics

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published