| --- |
| license: apache-2.0 |
| pipeline_tag: text-classification |
| tags: |
| - transformers |
| - text-classification |
| - multi-label |
| - energy |
| --- |
| |
| # Energy News Classifier |
|
|
| ## Overview |
|
|
| This model is a multi-label text classification system designed to extract structured signals from unstructured news data. |
|
|
| It focuses on identifying themes related to global energy systems, macroeconomic shifts, and geopolitical dynamics. |
|
|
| The model is built on top of DistilBERT and fine-tuned for domain-aware classification of news headlines and articles. |
|
|
| --- |
|
|
| ## Motivation |
|
|
| Energy is one of the most critical drivers of global systems. |
|
|
| Changes in supply chains, geopolitical tensions, regulation, and trade flows directly impact: |
|
|
| - Commodity markets |
| - Inflation cycles |
| - Global logistics |
| - Financial systems |
|
|
| Most traditional NLP models treat news as generic categories. This model instead focuses on extracting **signal-level intelligence** from news streams. |
|
|
| --- |
|
|
| ## Labels |
|
|
| The model supports multi-label classification across: |
|
|
| - energy |
| - politics |
| - trade |
| - stocks |
| - regulation |
| - shipping |
| - macro |
| - business |
| - technology |
| - risk |
|
|
| --- |
|
|
| ## Model Details |
|
|
| - Base Model: `distilbert-base-uncased` |
| - Task: Multi-label classification |
| - Framework: Hugging Face Transformers |
| - Output: Sigmoid probabilities |
|
|
| --- |
|
|
| ## Usage — Transformers (Recommended) |
|
|
| ```python |
| from transformers import pipeline |
| |
| classifier = pipeline( |
| "text-classification", |
| model="QuantBridge/energy-news-classifier", |
| top_k=None |
| ) |
| |
| classifier("Oil supply disrupted due to geopolitical tensions") |
