Skip to content

Sentiment analysis pipeline for commodity trading using DistilBERT, MiniLM, and GloVe-LSTM models. Processes real-time NewsData.io, MarketAux, Reddit data to predict 1-5 sentiment scores, mapping to Buy/Sell/Hold signals. Features oversampling, class-weighted loss, and backtesting with regression/classification metrics.

License

Notifications You must be signed in to change notification settings

khresth/NLP-Commodity-Signaller

Repository files navigation

NLP-Commodity-Signaller

Sentiment analysis pipeline for commodity trading using DistilBERT, MiniLM, and GloVe-LSTM models. Processes real-time NewsData.io, MarketAux, Reddit data to predict 1-5 sentiment scores, mapping to Buy/Sell/Hold signals. Features oversampling, class-weighted loss, and backtesting with regression/classification metrics.

Screenshot 2025-03-21 151146 Screenshot 2025-03-21 151211 Screenshot 2025-03-21 151317

About

Sentiment analysis pipeline for commodity trading using DistilBERT, MiniLM, and GloVe-LSTM models. Processes real-time NewsData.io, MarketAux, Reddit data to predict 1-5 sentiment scores, mapping to Buy/Sell/Hold signals. Features oversampling, class-weighted loss, and backtesting with regression/classification metrics.

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published