Leveraging Language Models for Carbon Market Insights: News Sentiment and Price Dynamics
Bibliographic record
Abstract
Abstract The carbon credit system plays a pivotal role in offsetting emissions, mitigating climate change, and enabling trading opportunities. We examine California's Low Carbon Fuel Standard (LCFS) using time series data from 2013 to 2024 to analyze carbon credit price dynamics and improve predictive capability with machine learning and large language models (LLMs). Technical analysis is employed to capture short-term trends (using monthly LCFS transaction data). While effective in identifying general price trends, these models struggle to adapt to shifts caused by policy changes or supply-demand fluctuations and offer limited insight into market dynamics. To address this, we incorporate news articles covering general carbon market topics. LLMs are employed for sentiment analysis, generating sentiment scores ranging from -1 (extremely negative) to 1 (extremely positive) and categorizing influence into short-term, mid-term, or long-term. The aggregated sentiment scores achieve over 60% alignment with price change. We further enhance prediction performance by integrating news data directly with trading data into advanced LLMs, including Gemini 1.5 Pro, Claude 3.5 Sonnet, GPT-4o, and o1-preview, resulting in higher F1 scores and improved accuracy. These LLMs demonstrated the ability to synthesize diverse information and provided clear market insights. For long-term forecasting, we integrate news data and LCFS trading data with California’s gasoline and diesel prices, annual CO2 emissions, electric vehicle sales, Cap-and-Trade (CaT) carbon tax prices, EU Emissions Trading Scheme (ETS) carbon prices, and Canada’s federal fuel charge into LLMs. The long-term prediction achieves F1 score up to 0.8, capturing price transitions and providing reasoned insights. This study highlights the potential of LLMs in carbon market forecasting, especially in enhancing interpretability and decision-making.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.001 | 0.004 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.001 |
| Bibliometrics | 0.002 | 0.001 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.001 | 0.001 |
| Open science | 0.000 | 0.000 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.002 | 0.001 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".