Pith. sign in

REVIEW 3 major objections 6 minor 36 references

Market-Information-Aware Gated-LoRA of Foundation Models for Transferable Day-Ahead Electricity Price Forecasting

T0 review · 3 major / 6 minor · reviewed 2026-08-15 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read Most of the transferable gain in day-ahead price forecasting comes from feeding a frozen foundation model the pre-clearing supply, demand, reserve, and intertie information; a six-parameter gate adds a smaller state-dependent gain.

desk verdict Careful LOMO study of transferable price forecasting; the MSMI gain hinges on an unvalidated version-policy assumption. read the letter →

arxiv 2608.11359 v1 pith:PH6RS477 submitted 2026-08-11 cs.LG

Hang Fan , Wei Wei , Shengwei Mei This is my paper · ORCID
classification cs.LG MSC 62M1062M2068T07
keywords electricitypriceforecastingtime-seriesfoundationmodeltransferlearninggatedLoRAday-aheadmarketleave-one-market-outevaluationprobabilisticinformationinterface
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

The paper tries to establish that day-ahead electricity price forecasting can be transferred to a market with no local training labels, and that the thing that makes transfer work is the information fed into the model, not the model itself. The authors feed the frozen Chronos-2 foundation model a 7-day price context plus the pre-clearing supply–demand, reserve, maintenance, generator-capacity, and intertie variables that market participants see before clearing, and fine-tune only about 1% of parameters as a low-rank adapter trained on related markets. Under a leave-one-market-out protocol on four Chinese provincial spot markets, the framework cuts average MAE/RMSE to 74.63/137.87, a 6.24%/7.99% improvement over the market-information-aware zero-shot backbone and a 3.05%/3.52% improvement over the ungated source adapter. The dominant gain is from the market-information interface itself; the state-dependent gate contributes a smaller, volatility-concentrated improvement that survives controls the authors run against global-scalar and shuffled-state alternatives. If correct, this gives newly established or data-scarce markets a deployment path that requires no local labels, and it redirects foundation-model work on market-cleared prices toward interface and adapter-control design rather than backbone redesign.

What carries the argument

Two objects carry the argument. The first is the multi-source market-information (MSMI) interface: a task interface that concatenates a 7-day (672-step) price context, its contemporaneous covariates, and the known future values of supply–demand, reserve, maintenance, generator-capacity, and intertie fields over the 1-day (96-step) forecast horizon, exposing the frozen backbone to the same pre-clearing information set a market participant holds. The second is the gated low-rank adapter: the LoRA update $W' = W + \gamma_\psi(z_t)\frac{\alpha}{r}BA$ is injected into the four attention projections and the patch-output projection of the frozen Chronos-2 weights, and a linear gate maps a five-dimensional standardized market-state vector $z_t$ — net load, renewable share, reserve adequacy, capacity margin, and recent price volatility, signed so that larger values mean tighter operating conditions — to a task-level multiplier clipped to $[1, 3]$. The gate is initialized from the source-trained LoRA checkpoint with only the reserve-tightness weight active, then trained for 300 source-market steps with the LoRA matrices frozen, adding six trainable parameters. The design's work is to let the source adapter strengthen exactly when the operating state tightens, without ever exposing a held-out market's labels.

What would settle it

Re-run the same leave-one-market-out protocol with the MSMI interface fed the final revised supply–demand rows instead of the earliest available ones, and again with the market-information variables lagged by one day so they could not be known pre-clearing: if either variant reproduces the 7.7% zero-shot MAE gain, the gain is not evidence of using the genuine pre-clearing information set. A complementary check is to deploy the framework on a day-ahead market that publishes only a subset of these fields before clearing; the MSMI advantage should shrink in proportion to the fields actually withheld.

Watch

Extended reading notes

Core claim

The paper's central claim is that a generic pretrained time-series model can forecast day-ahead electricity prices in a market it has never seen, provided the model is handed the economic information set from which those prices are formed. The authors construct a multi-source market-information (MSMI) interface — 672 hours of price history, contemporaneous covariates, and known day-ahead values of load, wind, solar, hydro, nuclear, reserve, ancillary-service, maintenance, generation-capacity, and intertie fields over the 96-step horizon — and show that switching Chronos-2's zero-shot input from a minimal load–wind–PV set to this interface reduces the four-market average MAE from 86.25 to 79.60 without any training. Training a rank-8 LoRA (about 1% of the 120.7M parameters) on the three source markets, with the backbone frozen and the held-out market untouched, brings MAE/RMSE to 76.98/142.90; replacing the adapter's fixed strength with a six-parameter gate that scales the LoRA update by reserve tightness, net load, renewable share, capacity margin, and recent volatility brings it to 74.63/137.87. The authors further claim, from control experiments with a learned global scalar, random gate initialization, and shuffled state–task pairing, that the gate's benefit is genuinely state-dependent rather than a generic calibration artifact. They are explicit that the gated gain is marginal outside the most volatile market and that it degrades 80% interval coverage, recommending a separate interval-calibration layer before operational use.

Load-bearing premise

The argument assumes the earliest available supply–demand table row is a truthful, revision-free stand-in for what the market knows before clearing; if load, renewable output, reserves, maintenance, and intertie figures are routinely revised after the fact or only published post-clearing, the large MSMI zero-shot gain would reflect information leakage rather than transferable forecasting skill.

Editorial extensions

If this is right

  • A newly established or data-scarce market can deploy day-ahead price forecasts immediately with zero local labels: the MSMI interface alone brings Chronos-2 zero-shot average MAE from 86.25 to 79.60.
  • Domain re-pretraining is unnecessary for this transfer: adapting about 1% of parameters on source markets reaches 76.98/142.90 average MAE/RMSE, and extending the source history from 900 to 1665 days changes results by only −0.24 MAE.
  • State-dependent adapter calibration works, but its benefit is concentrated: the reserve-initialized gate reaches 74.63/137.87, with the largest drop on the most volatile held-out market and near-zero change on calmer ones.
  • Probabilistic output is a separate axis: gated LoRA improves CRPS from 62.66 to 61.23 but drops pooled 80% interval coverage from 73.24% to 70.23%, so interval calibration needs an added layer before operational use.
  • Local few-shot labels are more useful after a source-domain warm start than from the pristine backbone: 30 target-market days give 76.66 average MAE from Source-LoRA versus 78.90 from pretrained Chronos-2.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • If the interface dominates adaptation, the same recipe likely transfers to other cleared prices with pre-market schedules — real-time and ancillary-service prices, emission allowances, or other auctioned commodities — where the pre-clearing information set is similarly published before the outcome is set.
  • The 'earliest version' supply–demand policy is only a proxy for the pre-clearing information set; a direct test is to feed the interface the final revised table rows instead. If the revised-version model outperforms the earliest-version model, later revisions carry predictive signal and part of the reported 7.7% interface gain is leakage rather than pre-clearing skill.
  • The gate's concentration in the volatile fold suggests a threshold hypothesis: the gate's benefit should appear mainly when reserve adequacy or capacity margin falls below a scarcity boundary, so a regime-stratified evaluation (tight vs slack days) would sharpen or refute the reserve-tightness story.
  • Concretely testable extension: the gate is six scalars; the paper's logic implies that a gate weighting only reserve tightness and net load — the two price-relevant scarcity signals — should match or beat the full five-dimensional gate on sparse data, indicating which market states actually carry transferable information.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

3 major / 6 minor

Summary. The paper addresses target-market-free transfer learning for day-ahead electricity price forecasting. It builds on Chronos-2, a pretrained time-series foundation model, and proposes (i) a multi-source market information (MSMI) interface that feeds seven days of price history plus forecast-horizon supply-demand, reserve, maintenance, capacity, and intertie variables into the frozen backbone; (ii) a source-domain LoRA adapter trained only on the three non-held-out markets under a leave-one-market-out (LOMO) protocol; and (iii) a market-state gate that scales the LoRA update by a linear function of five standardized state scores (net load, renewable share, reserve adequacy, capacity margin, volatility). On four Chinese provincial day-ahead markets, the MSMI interface reduces zero-shot MAE from 86.25 (Core) to 79.60, source-domain LoRA reduces it to 76.98, and the reserve-initialized gated LoRA achieves 74.63 MAE and 137.87 RMSE, a 3.05% MAE improvement over Source-LoRA. The authors also report CRPS, PICP, DM significance tests with HLN correction, an information-interface ablation, and a progressive few-shot adaptation spectrum. They conclude that the task interface is the dominant source of gain and that gated LoRA gives a small but useful additional calibration.

Significance. The paper is significant for the electricity price forecasting community because it tests a concrete, label-free transfer path for data-scarce markets and does so with unusual care for leakage: LOMO folds, a deterministic minimum-version data policy, source-only gate scaler statistics, and multiple control configurations for the gate. The finding that a broader pre-clearing information interface, rather than model-side changes, drives most of the improvement is a useful and falsifiable message. However, the headline numerical gains are only as credible as the assumption that the supply-demand table versions used are truly pre-clearing; the evidence for that assumption is currently absent, and the gated-LoRA advantage over Source-LoRA is modest and only marginally significant. With additional validation of the information timeline and more cautious language, this would be a solid contribution.

major comments (3)
  1. [III.A.1 and III.C.4] The central empirical claim rests on the assertion that the deterministic minimum-version policy 'approximates the information that would be available before day-ahead clearing.' The paper does not provide evidence that version 0 precedes the clearing deadline for any of the four markets, nor that the load, renewable, reserve, maintenance, capacity, and intertie fields are pre-clearing forecasts rather than settled or revised values. Because the MSMI-vs-Core improvement (86.25 to 79.60) and all subsequent Source-LoRA and Gated-LoRA gains are built on these inputs, hindsight leakage in these fields would convert the reported 'transferable' gains into artifacts. Please add release-schedule/timestamp evidence, a lagged-covariate control (feeding information as it would have been known at the forecast origin), or a comparison using only documented pre-clearing variables; without this, the main transferability claim is not established.
  2. [II.D, III.B.2, and Table IV] The evidence that state-dependent gating per se (rather than the initialization prior or small parameterization) explains the improvement is weak. The pooled Gated-vs-Source DM statistic is only marginally significant (p=0.074), Shanxi is not improved (statistic +0.49), and the shuffled-state control (74.873±0.047) is close to the matched reserve-initialized gate (74.633±0.002). The authors acknowledge the small matched-vs-shuffled gap in III.B.1 but elsewhere state that 'reserve-tightness-conditioned calibration can further improve' the adapter as a supported finding. The causal claim should be softened to 'the reserve-initialized gate is a stable, compact calibration that yields a small net improvement,' or additional evidence of state-dependence (e.g., market-level analysis showing larger gate multipliers on tight-reserve days are associated with lower loss) should be provided.
  3. [Abstract and III.B.3] The abstract and title foreground the gated-LoRA numbers (74.63/137.87) and the 3.05% reduction, while the body correctly describes the gain as marginally significant and largely limited to LN. This creates a mismatch between the strength of the evidence and the prominence of the claim. Please either report the gating result with explicit significance caveats in the abstract, or reposition the contribution as primarily the MSMI interface plus source-domain LoRA, with gating as a secondary calibration device. The conclusion should also state prominently that gated LoRA degrades interval calibration (PICP@80 falls to 70.23% and Macro-ACD rises to 9.77 p.p.).
minor comments (6)
  1. [II.C] The sentence 'These quantities are published or scheduled before market clearing' (the paragraph after Eq. (5)) is an unsupported assertion; please cite market documentation or replace it with a statement about the assumed information set.
  2. [III.B.3] Table V shows that gated LoRA lowers PICP@80 to 70.23% and increases Macro-ACD to 9.77 p.p.; the paper already notes this, but it should be reflected in the abstract's summary of the gating method.
  3. [III.B.6] The statement that 'representative windows are selected by transparent diagnostic rules' is not sufficient because the rules are not given; please specify them or state explicitly that the panels are illustrative only.
  4. [II.D and III.A.1] The gate uses beta=2.0 and clip range [1.0,3.0] in Eq. (12), but these values are introduced only in Section III.A.1; please define them where the gate is first presented.
  5. [Table III] The global-scalar row reports 77.001±0.000, which differs slightly from Source-LoRA (76.98); please explain whether this is a single run or multiple runs collapsed, since a zero standard deviation over three seeds is otherwise surprising.
  6. [III.A.1] The reported version-0 share (654377 out of 667680 rows) is useful, but the paper should also report the distribution of version numbers by market and by field; this would help readers assess the leakage risk.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: held-out target markets, source-only adapter/gate training, and external baselines keep the central claim self-contained.

full rationale

The paper's derivation chain is not circular. The central empirical claim is that a gated LoRA adapter on the MSMI interface reduces four-market average MAE/RMSE from 76.98/142.90 (Source-LoRA) to 74.63/137.87. The training objectives in Eqs. (11) and the gate objective optimize only on D_src, the source-market rolling tasks, and evaluation is on D_tar for a held-out market with no target-market labels used for training or model selection. The market-state scaler is standardized from source-market windows only, and the gate inputs (reserve tightness, net load, renewable share, capacity tightness, volatility) are exogenous scheduled/operating-state quantities, not the forecast target. No prediction is defined in terms of its input, and no fitted parameter is renamed as a forecast. The control experiments (global scalar, shuffled state, zero/random/reserve initialization) directly test whether the gate result is an artifact of construction. The paper invokes no self-citation as load-bearing evidence: references to Chronos-2, LoRA, and baseline models are external, and no uniqueness theorem from the authors is used to force a choice. The main assumption to scrutinize is the Section III.A.1 minimum-version policy, which assumes version-0 supply-demand rows approximate the pre-clearing information set; Section III.C.4 explicitly acknowledges that future market-information availability is a limitation. If the version policy is wrong, the risk is information leakage and inflated gains, not equivalence-by-construction or fitted-input circularity. Accordingly, no circular step can be exhibited, and the appropriate finding is a non-finding with score 0.

Assumptions & free parameters 4 free parameters · 6 assumptions · 0 invented entities

The paper introduces no new physical entity. The MSMI interface, market-state vector, and gated LoRA are software constructs built from known components. The main unvalidated assumptions concern the availability and timing of day-ahead market information, the comparability of the four selected markets, and the hand-set gate design.

free parameters (4)
  • LoRA rank and scaling = r=8, alpha=16
    Chosen hyperparameters for the source-domain adapter. They are not fitted to the target market, but the transfer result depends on this configuration.
  • Gate hyperparameters = beta=2, clip [1.0,3.0], linear gate, reserve weight initialized to 1 and others to 0
    Hand-set design choices for the gating mechanism. Controls show random and zero initializations do not reproduce the result, but the exact values are not derived.
  • Context length and horizon = L=672 and H=96 at 15-minute resolution
    Task definition fixed as 7-day context and 1-day horizon. A different context or horizon could change transfer gains.
  • Source training pool cap = 900 days per market
    Source history is capped; the authors verify that extending to full histories changes the average MAE by only -0.24, so this is not a load-bearing free parameter.
assumptions (6)
  • domain assumption Day-ahead market-information variables are known and correct over the forecast horizon at inference time.
    Section III.A.1 and III.C.1; if false, MSMI gains may reflect information leakage rather than transferable skill.
  • domain assumption The minimum-version row selection approximates the pre-clearing information set without hindsight leakage.
    Section III.A.1; later revisions are excluded, but the actual timing of data availability cannot be verified externally.
  • domain assumption The four selected Chinese provincial markets are comparable and representative enough for leave-one-market-out transfer conclusions.
    Section III.A.1; non-comparable records are excluded, which is a selection choice that affects the generality of the results.
  • standard math The Diebold-Mariano test with Newey-West HAC variance and Harvey-Leybourne-Newbold correction is valid for daily-window loss differentials.
    Section III.A.4; this is a standard statistical procedure, but its validity depends on the window-level loss series being well behaved.
  • ad hoc to paper A linear gate on five standardized market-state scores, clipped to [1.0,3.0], is a sufficient functional form for state-dependent adapter scaling.
    Section II.D, Equations (12)-(13); this functional form is a modeling choice rather than a derived consequence.
  • ad hoc to paper Reserve-initialized gate weights with all other weights zero provide the right inductive bias for market-state calibration.
    Section III.A.1; the initialization is motivated by reserve-tightness economics, but is not independently validated outside this dataset.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Market-Information-Aware Gated-LoRA of Foundation Models for Transferable Day-Ahead Electricity Price Forecasting." pith.science (2026). https://pith.science/paper/PH6RS477

@misc{pith2026260811359,
  author       = {Pith},
  title        = {Pith review of: Market-Information-Aware Gated-LoRA of Foundation Models for Transferable Day-Ahead Electricity Price Forecasting},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/PH6RS477}},
  note         = {Machine review of arXiv:2608.11359}
}
abstract

Electricity price forecasting is crucial for market participants but remains difficult because prices are volatile, market-specific, and closely tied to anticipated system conditions. Existing supervised methods depend largely on market-specific historical data, limiting their use in newly established or data-scarce markets. This paper proposes a market-information-aware adaptation framework that transfers the Chronos-2 time-series foundation model to day-ahead electricity price forecasting. It first constructs a multi-source market information (MSMI) interface aligning 7-day price context with pre-clearing supply--demand, reserve, maintenance, generator-capacity, and intertie variables, and then trains a source-domain gated low-rank adapter (LoRA), updating about $1\%$ of model parameters without target-market labels. The gate scales the frozen source adapter according to reserve-tightness and operating-state signals. A leave-one-market-out protocol is adopted for evaluating cross-market transferability. Experiments on four Chinese provincial day-ahead spot markets show that the proposed framework reduces the average MAE/RMSE by $6.24\%/7.99\%$ relative to market-information-aware zero-shot Chronos-2 and by $3.05\%/3.52\%$ relative to vanilla Source-LoRA. Experiments show that the gain is not reproduced by a learned global scalar or by random gate initialization, while the additional improvement over Source-LoRA is limited. These results suggest that market-structured inputs and state-dependent gated LoRA can provide a practical transfer path for data-scarce electricity markets.

Figures

Figures reproduced from arXiv: 2608.11359 by the authors.

Figure 1
Figure 1. An overview of the proposed method and testing windows are generated from S(m⋆ ) and m⋆ , respectively, thus target-market labels are not exposed. For each market m and forecast origin t, let y (m) t be the day-ahead clearing price. In the Core interface, the day-ahead market-information vector x (m) t ∈ R d contains load, wind gen￾eration, and photovoltaic generation. In the proposed MSMI interface, it is expanded … view at source ↗
Figure 2
Figure 2. Average day-ahead MAE comparison under LOMO evaluation. [PITH_FULL_IMAGE:figures/full_fig_p007_2.png] view at source ↗
Figure 4
Figure 4. Representative day-ahead prediction curves from real Chronos-2 [PITH_FULL_IMAGE:figures/full_fig_p008_4.png] view at source ↗
Figures from the paper (2 more)
Figure 3
Figure 3. Figure 3: Progressive target-market adaptation behavior under the MSMI [PITH_FULL_IMAGE:figures/full_fig_p008_3.png]
Figure 5
Figure 5. Figure 5: Per-window MAE distribution over 83 rolling day-ahead forecasts for [PITH_FULL_IMAGE:figures/full_fig_p009_5.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

36 extracted references · 27 canonical work pages

  1. [1]

    Electricity price forecasting: A review of the state-of-the- art with a look into the future,

    R. Weron, “Electricity price forecasting: A review of the state-of-the- art with a look into the future,”Int. J. Forecasting, vol. 30, no. 4, pp. 1030–1081, Oct. 2014

    work page 2014

  2. [2]

    Recent advances in electricity price forecasting: A review of probabilistic forecasting,

    J. Nowotarski and R. Weron, “Recent advances in electricity price forecasting: A review of probabilistic forecasting,”Renew. Sustain. Energy Rev., vol. 81, pp. 1548–1568, Jan. 2018

    work page 2018

  3. [3]

    Power market reform in China: Motivations, progress, and recommendations,

    H. Guo, M. R. Davidson, Q. Chen, D. Zhang, N. Jiang, Q. Xia, C. Kang, and X. Zhang, “Power market reform in China: Motivations, progress, and recommendations,”Energy Policy, vol. 145, p. 111717, Oct. 2020

    work page 2020

  4. [4]

    Forecasting day- ahead electricity prices: A review of state-of-the-art algorithms, best practices and an open-access benchmark,

    J. Lago, G. Marcjasz, B. De Schutter, and R. Weron, “Forecasting day- ahead electricity prices: A review of state-of-the-art algorithms, best practices and an open-access benchmark,”Appl. Energy, vol. 293, p. 116983, Jul. 2021

    work page 2021

  5. [5]

    Day-ahead electricity price forecasting with high-dimensional structures: Univariate vs. multivariate modeling frame- works,

    F. Ziel and R. Weron, “Day-ahead electricity price forecasting with high-dimensional structures: Univariate vs. multivariate modeling frame- works,”Energy Econ., vol. 70, pp. 396–420, Feb. 2018

    work page 2018

  6. [6]

    Putting big data analytics to work: Feature selection for forecasting electricity prices using the LASSO and random forests,

    N. Ludwig, S. Feuerriegel, and D. Neumann, “Putting big data analytics to work: Feature selection for forecasting electricity prices using the LASSO and random forests,”J. Decis. Syst., vol. 24, no. 1, pp. 19–36, Jan. 2015

    work page 2015

  7. [7]

    Variance stabilizing transforma- tions for electricity spot price forecasting,

    B. Uniejewski, R. Weron, and F. Ziel, “Variance stabilizing transforma- tions for electricity spot price forecasting,”IEEE Trans. Power Syst., vol. 33, no. 2, pp. 2219–2229, Mar. 2018

    work page 2018

  8. [8]

    Interpretable transformer model for capturing regime switching effects of real-time electricity prices,

    J. Bottieau, Y . Wang, Z. De Gr `eve, F. Vall ´ee, and J.-F. Toubeau, “Interpretable transformer model for capturing regime switching effects of real-time electricity prices,”IEEE Trans. Power Syst., vol. 38, no. 3, pp. 2162–2176, May 2023

    work page 2023

Show all 36 references
  1. [9]

    A novel electricity price forecasting approach based on dimension reduction strategy and rough artificial neural net- works,

    H. Jahangir, H. Tayarani, S. Baghali, A. Ahmadian, A. Elkamel, M. A. Golkar, and M. Castilla, “A novel electricity price forecasting approach based on dimension reduction strategy and rough artificial neural net- works,”IEEE Trans. Ind. Informat., vol. 16, no. 4, pp. 2369–2381...

    2020

  2. [10]

    Hybrid deep learning model for electricity price forecasting in renewable-dominated markets: The case of south australia,

    N. Vahedi, A. Jolfaei, S. U. Rehman, and R. S. Ravi, “Hybrid deep learning model for electricity price forecasting in renewable-dominated markets: The case of south australia,”IEEE Trans. Consumer Electron., vol. 72, no. 1, pp. 1227–1229, Feb. 2026

    2026

  3. [11]

    Temporal fusion trans- formers for interpretable multi-horizon time series forecasting,

    B. Lim, S. ¨O. Arik, N. Loeff, and T. Pfister, “Temporal fusion trans- formers for interpretable multi-horizon time series forecasting,”Int. J. Forecasting, vol. 37, no. 4, pp. 1748–1764, Oct. 2021

    2021

  4. [12]

    Autoformer: Decomposition transformers with auto-correlation for long-term series forecasting,

    H. Wu, J. Xu, J. Wang, and M. Long, “Autoformer: Decomposition transformers with auto-correlation for long-term series forecasting,” in Adv. Neural Inf. Process. Syst., vol. 34, Dec. 2021, pp. 22 419–22 430

    2021

  5. [13]

    FEDformer: Frequency enhanced decomposed transformer for long-term series forecasting,

    T. Zhou, Z. Ma, Q. Wen, X. Wang, L. Sun, and R. Jin, “FEDformer: Frequency enhanced decomposed transformer for long-term series forecasting,” inProc. 39th Int. Conf. Mach. Learn., ser. Proc. Mach. Learn. Res., vol. 162. PMLR, Jul. 2022, pp. 27 268–27 286. [Online]. Available: ...

    2022 arXiv

  6. [14]

    Probabilistic electricity price forecast with optimal prediction interval,

    C. Zhang and Y . Fu, “Probabilistic electricity price forecast with optimal prediction interval,”IEEE Trans. Power Syst., vol. 39, no. 1, pp. 442– 452, Jan. 2024

    2024

  7. [15]

    Conditional density forecast of electricity price based on ensemble ELM and logistic EMOS,

    S. Chai, Z. Xu, and Y . Jia, “Conditional density forecast of electricity price based on ensemble ELM and logistic EMOS,”IEEE Trans. Smart Grid, vol. 10, no. 3, pp. 3031–3043, May 2019

    2019

  8. [16]

    Predicting real-time locational marginal prices: A GAN-based approach,

    Z. Zhang and M. Wu, “Predicting real-time locational marginal prices: A GAN-based approach,”IEEE Trans. Power Syst., vol. 37, no. 2, pp. 1286–1296, Mar. 2022

    2022

  9. [17]

    Transfer learning for electricity price forecasting,

    S. Gunduz, U. Ugurlu, and I. Oksuz, “Transfer learning for electricity price forecasting,”Sustain. Energy Grids Netw., vol. 34, p. 100996, Jun. 2023

    2023

  10. [18]

    Transferable wind power probabilistic forecasting based on multi-domain adversarial networks,

    X. Dong, Y . Sun, L. Dong, J. Li, Y . Li, and L. Di, “Transferable wind power probabilistic forecasting based on multi-domain adversarial networks,”Energy, vol. 285, p. 129496, Dec. 2023

    2023

  11. [19]

    Chronos-2: From univariate to universal forecasting,

    A. F. Ansari, O. Shchur, J. K ¨uken, A. Auer, B. Han, P. Mercado, S. S. Rangapuram, H. Shen, L. Stella, X. Zhang, M. Goswami, S. Kapoor, D. C. Maddix, P. Guerron, T. Hu, J. Yin, N. Erickson, P. M. Desai, H. Wang, H. Rangwala, G. Karypis, Y . Wang, and M. Bohlke- Schneider, “Ch...

    arXiv

  12. [20]

    Lag-Llama: Towards foundation models for probabilistic time series forecasting,

    K. Rasul, A. Ashok, A. R. Williams, H. Ghonia, R. Bhagwatkar, A. Khorasani, M. J. D. Bayazi, G. Adamopoulos, R. Riachi, N. Hassen, M. Bilo ˇs, S. Garg, A. Schneider, N. Chapados, A. Drouin, V . Zantedeschi, Y . Nevmyvaka, and I. Rish, “Lag-Llama: Towards foundation models for ...

    arXiv

  13. [21]

    Unified training of universal time series forecasting transformers,

    G. Woo, C. Liu, A. Kumar, C. Xiong, S. Savarese, and D. Sahoo, “Unified training of universal time series forecasting transformers,” in Proc. 41st Int. Conf. Mach. Learn., ser. Proc. Mach. Learn. Res., vol

  14. [22]

    A decoder-only foundation model for time-series forecasting,

    A. Das, W. Kong, R. Sen, and Y . Zhou, “A decoder-only foundation model for time-series forecasting,” inProc. 41st Int. Conf. Mach. Learn., ser. Proc. Mach. Learn. Res., vol. 235. PMLR, Jul. 2024, pp. 10 148–10 167. [Online]. Available: https: //proceedings.mlr.press/v235/das24c.html

    2024

  15. [23]

    TimeXer: Empowering transformers for time series forecasting with exogenous variables,

    Y . Wang, H. Wu, J. Dong, G. Qin, H. Zhang, Y . Liu, Y . Qiu, J. Wang, and M. Long, “TimeXer: Empowering transformers for time series forecasting with exogenous variables,” inAdv. Neural Inf. Process. Syst., vol. 37, Dec. 2024, pp. 469–498. [Online]. Available: https://proceed...

    2024

  16. [24]

    PriceFM: Foundation model for probabilistic electricity price forecasting,

    R. Yu, C. Gu, J. Stiasny, Q. Wen, W. S. Dilov, L. Qi, and J. L. Cremer, “PriceFM: Foundation model for probabilistic electricity price forecasting,”arXiv preprint arXiv:2508.04875, 2025. [Online]. Available: https://arxiv.org/abs/2508.04875

    2025 arXiv

  17. [25]

    Wind power forecasting based on large time series model,

    J. Yan, Y . Li, H. Wang, S. Han, W. Shang, and Y . Liu, “Wind power forecasting based on large time series model,”Engineering, Nov. 2025, available online 19 Nov. 2025

    2025

  18. [26]

    Chronos: Learning the language of time series,

    A. F. Ansari, L. Stella, A. C. Turkmen, X. Zhang, P. Mercado, H. Shen, O. Shchur, S. S. Rangapuram, S. Pineda Arango, S. Kapoor, J. Zschiegner, D. C. Maddix, H. Wang, M. W. Mahoney, K. Torkkola, A. G. Wilson, M. Bohlke-Schneider, and B. Wang, “Chronos: Learning the language of...

    2024

  19. [27]

    Strictly proper scoring rules, prediction, and estimation,

    T. Gneiting and A. E. Raftery, “Strictly proper scoring rules, prediction, and estimation,”J. Amer. Statist. Assoc., vol. 102, no. 477, pp. 359–378, Mar. 2007

    2007

  20. [28]

    LoRA: Low-rank adaptation of large language models,

    E. J. Hu, Y . Shen, P. Wallis, Z. Allen-Zhu, Y . Li, S. Wang, L. Wang, and W. Chen, “LoRA: Low-rank adaptation of large language models,” inProc. Int. Conf. Learn. Represent., 2022. [Online]. Available: https://openreview.net/forum?id=nZeVKeeFYf9

    2022

  21. [29]

    XGBoost: A scalable tree boosting system,

    T. Chen and C. Guestrin, “XGBoost: A scalable tree boosting system,” inProc. 22nd ACM SIGKDD Int. Conf. Knowl. Discovery Data Mining, Aug. 2016, pp. 785–794

    2016

  22. [30]

    LightGBM: A highly efficient gradient boosting decision tree,

    G. Ke, Q. Meng, T. Finley, T. Wang, W. Chen, W. Ma, Q. Ye, and T.-Y . Liu, “LightGBM: A highly efficient gradient boosting decision tree,” in Adv. Neural Inf. Process. Syst., vol. 30, Dec. 2017

    2017

  23. [31]

    A time series is worth 64 words: Long-term forecasting with transformers,

    Y . Nie, N. H. Nguyen, P. Sinthong, and J. Kalagnanam, “A time series is worth 64 words: Long-term forecasting with transformers,” inProc. Int. Conf. Learn. Represent., 2023. [Online]. Available: https://arxiv.org/abs/2211.14730

    2023 arXiv

  24. [32]

    iTransformer: Inverted transformers are effective for time series forecasting,

    Y . Liu, T. Hu, H. Zhang, H. Wu, S. Wang, L. Ma, and M. Long, “iTransformer: Inverted transformers are effective for time series forecasting,” inProc. Int. Conf. Learn. Represent., 2024. [Online]. Available: https://arxiv.org/abs/2310.06625

    2024 arXiv

  25. [33]

    Are transformers effective for time series forecasting?

    A. Zeng, M. Chen, L. Zhang, and Q. Xu, “Are transformers effective for time series forecasting?” inProc. AAAI Conf. Artif. Intell., vol. 37, no. 9, Jun. 2023, pp. 11 121–11 128. [Online]. Available: https://arxiv.org/abs/2205.13504

    2023 arXiv

  26. [34]

    Comparing predictive accuracy,

    F. X. Diebold and R. S. Mariano, “Comparing predictive accuracy,”J. Bus. Econ. Statist., vol. 13, no. 3, pp. 253–263, Jul. 1995

    1995

  27. [35]

    Testing the equality of prediction mean squared errors,

    D. Harvey, S. Leybourne, and P. Newbold, “Testing the equality of prediction mean squared errors,”Int. J. Forecasting, vol. 13, no. 2, pp. 281–291, Jun. 1997

    1997

  28. [235]

    2024, pp

    PMLR, Jul. 2024, pp. 53 140–53 164

    2024

Pith tools

Reviewed August 15, 2026 · model on record in the stance chip above.