Comparative Analysis of Fine-Tuning Time Series Foundation Models in Short-Term Load Prediction
Bibliographic record
Abstract
Large Language Models, or LLMs, have transformed natural language processing and are now being adapted for non-textual tasks, including time series forecasting. This paper explores the fine-tuning of LLM-based and LLM-inspired Time Series Foundation Models, known as TSFMs, for short-term electric load prediction in a regional context. Using historical load data from the Alberta Electric System Operator spanning 2011 to 2023, we evaluate five recent TSFMs: MOMENT, PatchTST, Tiny Time Mixers (TTMixer), TimeLLM, and AutoTimes. While TimeLLM and AutoTimes explicitly repurpose LLMs for timeseries modeling, models like MOMENT and PatchTST apply similar transformer-based architectures. Our findings show that out-of-the-box TSFMs underperform due to the complex and evolving nature of Alberta’s electricity consumption, but finetuning yields significant improvements. Our results show that medium-sized models, such as TTMixer and TimeLLM outperform larger architectures, achieving high forecasting accuracy and robust uncertainty estimation.
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 | 0.003 | 0.010 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.001 | 0.000 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.001 | 0.001 |
| Open science | 0.001 | 0.001 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.001 | 0.000 |
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".