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.
khresth/NLP-Commodity-Signaller
Folders and files
| Name | Name | Last commit date | ||
|---|---|---|---|---|


