Mixer-Informer-Based Two-Stage Transfer Learning for Long-Sequence Load Forecasting in Newly Constructed Electric Vehicle Charging Stations

Bozhen Jiang; Qin Wang; Zhenhua Zhou

REVIEW 4 major objections 5 minor 27 references

Mixer-Informer-Based Two-Stage Transfer Learning for Long-Sequence Load Forecasting in Newly Constructed Electric Vehicle Charging Stations

T0 review · 4 major / 5 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 MIK-TST reduces EV charging load forecast error by 4% in MAE and 8% in MSE relative to the best baseline.

desk verdict Uncontrolled comparison: without a transfer-learning baseline, the 4/8% gains in MIK-TST likely come from extra source data, not the architecture. read the letter →

arxiv 2505.06657 v1 pith:C6SMSXYL submitted 2025-05-10 eess.SY cs.SY

Zhenhua Zhou , Bozhen Jiang , Qin Wang This is my paper · ORCID

classification eess.SYcs.SY

keywords EVchargingloadforecastingtransferlearningMixerInformerKANlong-sequencetime-seriesProbSparseattentionfine-tuning

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

This paper proposes MIK-TST, a forecasting model that combines a Mixer feature-fusion layer, an Informer with ProbSparse attention, and a Kolmogorov-Arnold Network, and trains it in two stages: pre-training on 21 established charging stations and fine-tuning on limited early data from five newer stations. The authors claim this setup reduces mean absolute error by 4% and mean squared error by 8% relative to the best baseline, PatchTST, on data from 26 EV charging stations in Boulder, USA. If true, the value is practical: newly built charging stations could get usable load forecasts from weeks of data instead of months or years, supporting grid planning and charging infrastructure expansion. The paper also reports ablation results attributing the gain chiefly to the transfer-learning stage, with the Mixer and KAN modules contributing smaller improvements.

What carries the argument

The load-bearing mechanism is the MIK-TST pipeline itself, which stacks three components: the Mixer module, an all-MLP structure with separate temporal-mixing and feature-mixing layers that fuse multi-source inputs; the Informer encoder-decoder, whose ProbSparse attention selects dominant queries to capture long-range dependencies at sub-quadratic cost; and a KAN module, which replaces fixed node activations with learnable spline functions on the edges. These are tied together by a two-stage transfer learning strategy: pre-train on complete histories of 21 source stations with MSE loss, then fine-tune on limited early data from five target stations, freezing Mixer and Informer weights and updating only KAN, or fine-tuning all parameters at a reduced learning rate. The framework's job is to make forecasts accurate enough for newly constructed stations where only weeks of data exist.

What would settle it

Run PatchTST or Informer through the same two-stage pre-train/fine-tune protocol on the same 21 source and 5 target Boulder stations; if either reaches or beats MAE 0.4450 and MSE 0.6322, the claimed architectural advantage of MIK-TST is not supported.

Watch

Extended reading notes

Core claim

On the paper's own terms, the central discovery is that a hybrid architecture plus a two-stage transfer learning protocol produces the best reported forecasts on the Boulder 26-station dataset: MIK-TST reaches MAE 0.4450 and MSE 0.6322, compared with PatchTST's 0.4638 and 0.7409, the best baseline on the two metrics. The authors interpret the gain as coming from the interaction of the three modules—Mixer for fusing temporal and channel features, Informer's ProbSparse attention for long-range dependencies, and KAN's learnable activations for nonlinear patterns—together with pre-training on 21 source stations before fine-tuning on five target stations. The ablation study supports the transfer stage as the largest contributor: removing transfer learning raises MAE by 7.2% and MSE by 18%, while removing Mixer raises MAE by 3% and MSE by 3.5%, and removing KAN raises MAE by 1.3% with negligible MSE change.

Load-bearing premise

The fairness of comparing MIK-TST, pre-trained on 21 source stations, against baselines trained only on the five target stations is the load-bearing premise; if that comparison is unfair, the claimed 4% and 8% gains may be a data effect, not an architecture effect.

Editorial extensions

If this is right

A newly opened charging station could receive a usable forecasting model after fine-tuning on a short history (the paper uses pre-2023 data), rather than waiting months to accumulate training data.
The ablation numbers imply the biggest single return comes from the transfer protocol, not from any one network module: without it, MSE rises 18%.
Hyperparameter settings that balance capacity against sparse fine-tuning data (embedding 256, 8 attention heads, 2 encoder/decoder layers) are part of the claimed performance, not incidental choices.
Because MSE improves more than MAE, the model reduces large forecasting errors disproportionately, which matters for grid operators worried about peak misprediction.

Reading between the lines

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

The reported comparison may overstate the architecture's contribution: baselines are trained only on target stations, so part of the 4%/8% edge could be the 21-station pre-training corpus. A transfer-enabled PatchTST or Informer baseline would separate data advantage from model advantage.
The KAN module's contribution looks small in the ablation (1.3% MAE, no MSE change), suggesting a simpler nonlinear head might perform almost as well; testing that substitution would clarify whether KAN earns its complexity.
With a single-city, 26-station dataset and station assignment by chronological opening order, the results may not generalize to stations with very different usage profiles; a multi-city or heterogeneous-station test would be more informative.
The paper's own limitations note the absence of weather, traffic, and socio-economic covariates; adding them could shift which module matters most and could make the transfer gains larger or smaller.

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.

Desk Editor's Note

read the letter

Colleague,

You should know: the paper's headline claim — 4% MAE and 8% MSE gains from the Mixer-Informer-KAN + transfer learning stack — is not established by the experiments. The architecture itself is an incremental assembly of three published modules with a standard pre-train/fine-tune loop. What's genuinely nice is the problem: newly built EV charging stations with little data are a real forecasting pain point, and the split by chronological station order is a sensible way to simulate a new station.

The paper does a few things right. It compares against a reasonable set of baselines (PatchTST, Autoformer, DLinear, Crossformer, FreTS). The ablation study explicitly shows how much transfer learning contributes. And the writing is honest about limitations. That matters.

But the central comparison is uncontrolled. MIK-TST is pre-trained on 21 source stations and then fine-tuned; every baseline is trained only on the five target stations. So the 4/8% could simply be the extra source data — and the ablation confirms it: removing transfer learning (w/o-Dcl) gives MAE 0.4772 / MSE 0.7457, worse than every baseline in Table 1. Without a transfer-learning-enabled baseline (e.g., a fine-tuned PatchTST or Informer), you cannot attribute the gains to the MIK architecture. This is a load-bearing flaw.

Other soft spots: single runs, no standard deviations, no significance tests. Hyperparameters were chosen on the same evaluation set used to report the headline numbers. There are also mechanical issues: Eq. 7 is vanilla softmax attention, not ProbSparse; the ablation labels (Mcl/Kcl/Dcl) are unexplained; and the parameter sensitivity section reports a single run per setting.

I want to be fair: none of this is fraudulent. The gains may even survive a cleaner test. But as it stands, the evidence is too thin for an applied forecasting paper. The authors need to add transfer baselines, report multiple seeds, and use a validation split for hyperparameter selection.

Who is this for? Applied researchers in EV charging / smart grid who want a transfer-learning template. With proper controls, it could be a decent niche study. I'd give them a shot in peer review — but only because the fix is clear and the problem is real. If the revision doesn't add those baselines, reject.

Referee Report

4 major / 5 minor

Summary. This paper proposes MIK-TST, a hybrid architecture that combines MLP-Mixer, Informer, and Kolmogorov-Arnold Networks with a two-stage transfer learning strategy for long-sequence load forecasting at newly constructed EV charging stations. The authors pre-train on 21 Boulder stations and fine-tune on early data from five target stations, reporting MAE 0.4450 and MSE 0.6322, interpreted as 4% and 8% improvements over baselines. An ablation study removes the Mixer, KAN, and transfer learning components, and a sensitivity analysis varies hidden dimension, attention heads, and layer depth.

Significance. The problem is practically relevant: newly constructed EV stations have scarce data, and transfer learning is a natural remedy. If the reported gains were genuine, this would be a useful contribution to the forecasting literature. However, the paper does not deliver a controlled empirical comparison: all baselines are trained without access to the source-domain data, and the sole ablation that removes transfer learning makes the model worse than every baseline, showing the reported gains come from additional data rather than from the MIK architecture. The paper also lacks standard deviations, multiple seeds, and statistical tests, and the hyperparameters of the proposed model are selected on the evaluation set. The central claim is therefore not established.

major comments (4)

[§4.4, Table 1 and §4.1] The comparison is not controlled. MIK-TST is pre-trained on 21 source stations and then fine-tuned on the target stations, while every baseline is trained only on the early target data. The paper never includes a transfer-learning-enabled baseline, such as a fine-tuned PatchTST or Informer. Consequently, the reported 4% MAE and 8% MSE gains cannot be attributed to the Mixer-Informer-KAN architecture. The ablation w/o-Dcl in Table 2 removes the two-stage transfer learning strategy and yields MAE 0.4772 and MSE 0.7457, which is worse than every baseline in Table 1; the authors even state this is a 7.2% MAE and 18% MSE increase. This demonstrates that the architecture alone is not superior to the baselines and that the headline improvement is driven by extra source-domain data.
[§4.3 and §4.6] Hyperparameter selection is performed on the evaluation set, introducing selection bias. The paper reports the best configuration among hidden dimension d in {64, 128, 256, 512, 1024}, attention heads in {1, 2, 4, 6, 8, 10}, and layers in {1, 2, 4, 6} without a held-out validation split. Baseline models are said to use original parameters without any tuning. This makes the comparison unfair and the reported improvements not statistically meaningful. Please provide a proper validation split, report performance for all configurations, and give standard deviations over multiple seeds.
[§4.4, Table 1] The claim that MIK-TST outperforms 'the best baseline' by 4% in MAE and 8% in MSE is internally inconsistent. The best MAE baseline is PatchTST (0.4638), but the best MSE baseline is Autoformer (0.6869); the 8% MSE improvement is relative to Autoformer, not PatchTST. The paper should state the per-metric baseline and report confidence intervals for both metrics.
[§4.1] The experimental protocol is underspecified. The paper does not provide the prediction horizon (e.g., 24 hours, 7 days), the backtest window, how 'early data prior to 2023' is split into training and validation, or the number of samples in the source and target domains. Without this information, the reader cannot reconstruct the experiment or assess the impact of the fine-tuning sample size. Please include a data description table and a precise train/validation/test split.

minor comments (5)

[Eq. (7)] The square root in the attention formula is rendered as the literal string 'sqrt(dk)' in the text and should be typeset as sqrt(d_k).
[§3.1.1, Eqs. (4)–(6)] The notation is garbled: x ∈ R^n and y ∈ R^m appear in Eq. (3), but the dimension compatibility with the mixing layers is not defined. The regularization function denoted by the symbol that resembles a script 'D' is introduced only as 'an effective regularization technique' and should be explicitly identified.
[§3.1.1 and throughout] There are multiple typographical errors, including 'TBy leveraging this mechanism' on the Mixer page, and inconsistent spacing in equations such as '𝑓𝐿(𝑋) = 𝑊𝑋 +𝑏'. Please proofread the text.
[Figs. 5–7] The parameter sensitivity figures are not described in the caption beyond the parameter name; they should include axis labels, units, and the range of the y-axis to be interpretable.
[Reproducibility] No code or data availability statement is provided, which limits the usefulness of the empirical results for the community.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the paper presents an empirical benchmark rather than a derivation that reduces to its inputs.

full rationale

The paper makes no derivation claim that reduces to its own inputs. The central result is an empirical comparison (Section 4.4, Table 1) in which MIK-TST is pre-trained on 21 source stations, fine-tuned on early target-station data, and evaluated on later target-station data. The Mixer, Informer, and KAN components are standard externally cited modules (Mixer [14], Informer [7], KAN [13]); none is defined in terms of the outcome metrics, and no uniqueness theorem or load-bearing self-citation is invoked. The ablation study in Table 2 compares model variants and does not rename a fitted parameter as a prediction. The main caveats are experimental-design issues rather than circularity: the baselines are not given the same transfer-learning treatment, the hyperparameters appear to be selected with knowledge of the reported performance, and the claimed 'best baseline' differs per metric. These concern fairness and reporting accuracy, not equivalence of the result to its inputs by construction. Since no quoted equation or step exhibits the required reduction, the circularity score is 0.

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

The central claim depends on three test-set-selected hyperparameters and on the assumption that pre-training on 21 stations transfers to 5 new stations. No new entities are introduced; the architecture concatenates three published modules.

free parameters (4)

hidden state dimension d = 256
Selected via test-set sensitivity scan (Fig. 5); used for final results.
number of attention heads N_heads = 8
Selected via test-set scan (Fig. 6).
number of encoder/decoder layers r = 2
Selected via test-set scan (Fig. 7).
learning rate = 1e-4
Fixed by hand in Section 4.3; not tuned but chosen.

assumptions (4)

standard math Kolmogorov-Arnold representation theorem
Used to justify KAN module in Section 3.3; imported from reference [13].
domain assumption Pre-training on stations 0-20 transfers useful patterns to stations 21-25
Core premise of the two-stage transfer strategy (Section 3.4); no experiment isolates positive transfer from architecture effects.
domain assumption Charging load is predictable from historical load and timestamp features
Problem formulation in Section 2.1; no analysis of predictability limits or upper bounds.
domain assumption The Boulder dataset is representative of newly constructed EV stations
All experiments use this single dataset; no cross-city validation.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Mixer-Informer-Based Two-Stage Transfer Learning for Long-Sequence Load Forecasting in Newly Constructed Electric Vehicle Charging Stations." pith.science (2026). https://pith.science/paper/C6SMSXYL

@misc{pith2026250506657,
  author       = {Pith},
  title        = {Pith review of: Mixer-Informer-Based Two-Stage Transfer Learning for Long-Sequence Load Forecasting in Newly Constructed Electric Vehicle Charging Stations},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/C6SMSXYL}},
  note         = {Machine review of arXiv:2505.06657}
}

read the original abstract

The rapid rise in electric vehicle (EV) adoption demands precise charging station load forecasting, challenged by long-sequence temporal dependencies and limited data in new facilities. This study proposes MIK-TST, a novel two-stage transfer learning framework integrating Mixer, Informer, and Kolmogorov-Arnold Networks (KAN). The Mixer fuses multi-source features, Informer captures long-range dependencies via ProbSparse attention, and KAN enhances nonlinear modeling with learnable activation functions. Pre-trained on extensive data and fine-tuned on limited target data, MIK-TST achieves 4% and 8% reductions in MAE and MSE, respectively, outperforming baselines on a dataset of 26 charging stations in Boulder, USA. This scalable solution enhances smart grid efficiency and supports sustainable EV infrastructure expansion.

Figures

Figures reproduced from arXiv: 2505.06657 by the authors.

**Figure 1.** Overall architecture of the MIK-TST However, these models often lack tailored transfer learning strategies, making them less effective for newly constructed charging stations where historical data is limited. This gap underscores the need for a hybrid framework like MIKTST, which combines feature fusion, long-range dependency modeling, and transfer learning to address the unique challenges of EV charging load fore… view at source ↗

**Figure 2.** Overall architecture of the Mixer • It is proposed to capture long-range dependencies in the time series, enabling the model to effectively utilize historical information from distant past observations. Feature/Channel Mixing Layer: • Following temporal mixing, the feature mixing layer applies a linear projection and fully connected operations across different feature channels. • This layer facilitates interaction… view at source ↗

**Figure 3.** Overall architecture of the Informer dimension of the key vectors. ProbSparse attention reduces unnecessary computation by probabilistically selecting key queries, thereby enabling effective processing of long sequences. The overall architecture of the Informer model follows the classic encoder-decoder design: • Encoder: Composed of multiple stacked attention layers that progressively extract high-level feature rep… view at source ↗

Figures from the paper (4 more)

**Figure 4.** Figure 4: Overall architecture of the KAN such long sequences [20]. The Informer model effectively overcomes these limitations by leveraging its ProbSparse attention mechanism [21]. This approach selectively focuses on the most relevant key queries, significantly reducing compu…

**Figure 5.** Figure 5: Parameter sensitivity analysis of MIK-TST with respect to the hidden state dimension 𝑑 [PITH_FULL_IMAGE:figures/full_fig_p008_5.png]

**Figure 6.** Figure 6: Parameter sensitivity analysis of MIK-TST with respect to the number of attention heads 𝑁ℎ𝑒𝑎𝑑𝑠. the KAN module (w/o-Kcl) raises MAE to 0.4508 and MSE to 0.6338, indicating a modest 1.3% MAE increase but minimal MSE impact. This suggests KAN’s learnable activation funct…

**Figure 7.** Figure 7: Parameter sensitivity analysis of MIK-TST with respect to the number of encoder and decoder layers 𝑟 narrow subset of dependencies, missing diverse patterns like peak-hour surges or weekend trends. Performance improves with more heads, peaking at 𝑁heads = 8 (MAE: 0.445…

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

27 extracted references · 22 canonical work pages

[1]

Global EV outlook 2024,

“Global EV outlook 2024,” International Energy Agency (IEA), Tech. Rep., 2024, Available: https://www.iea.org/reports/global-ev- outlook-2024

work page 2024
[2]

Electric vehicle charging load forecasting: A comparative studyofdeeplearningapproaches,

J.Zhu,Z.Yang,M.Mourshed,Y.Guo,Y.Zhou,Y.Chang,Y.Wei,and S. Feng, “Electric vehicle charging load forecasting: A comparative studyofdeeplearningapproaches,” Energies,vol.12,no.14,p.2692, 2019

work page 2019
[3]

Elec- tric vehicle charging load forecasting considering weather impact,

W. Wang, A. Tang, F. Wei, H. Yang, L. Xinran, and J. Peng, “Elec- tric vehicle charging load forecasting considering weather impact,” Applied Energy, vol. 383, p. 125337, 2025

work page 2025
[4]

A transfer learning method for electric vehicles charging strategy based ondeepreinforcementlearning,

K. Wang, H. Wang, Z. Yang, J. Feng, Y. Li, J. Yang, and Z. Chen, “A transfer learning method for electric vehicles charging strategy based ondeepreinforcementlearning,” AppliedEnergy,vol.343,p.121186, 2023

work page 2023
[5]

Review on schedul- ing,clustering,andforecastingstrategiesforcontrollingelectricvehi- cle charging: Challenges and recommendations,

A. S. Al-Ogaili, T. J. T. Hashim, N. A. Rahmat, A. K. Ramasamy, M. B. Marsadek, M. Faisal, and M. A. Hannan, “Review on schedul- ing,clustering,andforecastingstrategiesforcontrollingelectricvehi- cle charging: Challenges and recommendations,”Ieee Access, vol. 7, pp. 128353–128371, 2019

work page 2019
[6]

R. H. Shumway, D. S. Stoffer, and D. S. Stoffer,Time series analysis and its applications. Springer, 2000, vol. 3

work page 2000
[7]

Informer:Beyondefficienttransformerforlongsequencetime-series forecasting,

H.Zhou,S.Zhang,J.Peng,S.Zhang,J.Li,H.Xiong,andW.Zhang, “Informer:Beyondefficienttransformerforlongsequencetime-series forecasting,” in Proceedings of the AAAI conference on artificial intelligence, vol. 35, no. 12, 2021, pp. 11106–11115

work page 2021
[8]

Atimeseries is worth 64 words: Long-term forecasting with transformers,

Y.Nie,N.H.Nguyen,P.Sinthong,andJ.Kalagnanam,“Atimeseries is worth 64 words: Long-term forecasting with transformers,”arXiv preprint arXiv:2211.14730, 2022

arXiv 2022

Show all 27 references

[9]

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,” Advances in neural information processing systems, vol. 34, pp. 22419–22430, 2021

2021
[10]

Aretransformerseffective fortimeseriesforecasting?

M.Chen,Q.Xu,A.Zeng,L.Zhang etal.,“Aretransformerseffective fortimeseriesforecasting?” arXivpreprint,arXiv:2205.13504 ,2022

2022 arXiv
[11]

Crossformer: Transformer utilizing cross- dimension dependency for multivariate time series forecasting,

Y. Zhang and J. Yan, “Crossformer: Transformer utilizing cross- dimension dependency for multivariate time series forecasting,” in The eleventh international conference on learning representations, 2023

2023
[12]

Frequency-domain mlps are more effective learners in time series forecasting,

K. Yi, Q. Zhang, W. Fan, S. Wang, P. Wang, H. He, N. An, D. Lian, L. Cao, and Z. Niu, “Frequency-domain mlps are more effective learners in time series forecasting,”Advances in Neural Information Processing Systems, vol. 36, pp. 76656–76679, 2023

2023
[13]

Kan: Kolmogorov-arnold networks,

Z.Liu,Y.Wang,S.Vaidya,F.Ruehle,J.Halverson,M.Soljačić,T.Y. Hou, and M. Tegmark, “Kan: Kolmogorov-arnold networks,”arXiv preprint arXiv:2404.19756, 2024

2024 arXiv
[14]

Mlp-mixer: An all-mlp architecture for vision,

I. O. Tolstikhin, N. Houlsby, A. Kolesnikov, L. Beyer, X. Zhai, T. Unterthiner, J. Yung, A. Steiner, D. Keysers, J. Uszkoreitet al., “Mlp-mixer: An all-mlp architecture for vision,”Advances in neural information processing systems, vol. 34, pp. 24261–24272, 2021

2021
[15]

Transfer learning- based framework enhanced by deep generative model for cold-start forecasting of residential ev charging behavior,

A. Forootani, M. Rastegar, and H. Zareipour, “Transfer learning- based framework enhanced by deep generative model for cold-start forecasting of residential ev charging behavior,”IEEE Transactions on Intelligent Vehicles, vol. 9, no. 1, pp. 190–198, 2023

2023
[16]

Improving neural networks by preventing co- adaptation of feature detectors,

G. E. Hinton, N. Srivastava, A. Krizhevsky, I. Sutskever, and R. R. Salakhutdinov, “Improving neural networks by preventing co- adaptation of feature detectors,” arXiv preprint arXiv:1207.0580, 2012

2012 arXiv
[17]

Deep sparse rectifier neural networks,

X. Glorot, A. Bordes, and Y. Bengio, “Deep sparse rectifier neural networks,” inProceedings of the fourteenth international conference on artificial intelligence and statistics. JMLR Workshop and Conference Proceedings, 2011, pp. 315–323

2011
[18]

Multimodal machine learning: A survey and taxonomy,

T. Baltrušaitis, C. Ahuja, and L.-P. Morency, “Multimodal machine learning: A survey and taxonomy,”IEEE transactions on pattern analysis and machine intelligence, vol. 41, no. 2, pp. 423–443, 2018

2018
[19]

Modeling long-and short-term temporal patterns with deep neural networks,

G. Lai, W.-C. Chang, Y. Yang, and H. Liu, “Modeling long-and short-term temporal patterns with deep neural networks,” inThe 41st international ACM SIGIR conference on research & development in information retrieval, 2018, pp. 95–104

2018
[20]

In- former model with season-aware block for efficient long-term power time series forecasting,

Y. Cui, Z. Li, Y. Wang, D. Dong, C. Gu, X. Lou, and P. Zhang, “In- former model with season-aware block for efficient long-term power time series forecasting,”Computers and Electrical Engineering, vol. 119, p. 109492, 2024

2024
[21]

Multimodaljoint prediction of traffic spatial-temporal data with graph sparse attention mechanism and bidirectional temporal convolutional network,

D.Zhang,J.Yan,K.Polat,A.Alhudhaif,andJ.Li,“Multimodaljoint prediction of traffic spatial-temporal data with graph sparse attention mechanism and bidirectional temporal convolutional network,”Ad- vanced Engineering Informatics, vol. 62, p. 102533, 2024

2024
[22]

Deep learning models for time series forecast- ing: a review,

W. Li and K. E. Law, “Deep learning models for time series forecast- ing: a review,”IEEE Access, 2024

2024
[23]

Enhancing neural function approxima- tion:Thexnetoutperformingkan,

X. Li, X. Zheng, and Z. Xia, “Enhancing neural function approxima- tion:Thexnetoutperformingkan,” arXivpreprintarXiv:2501.18959 , 2025

2025 arXiv
[24]

An improvised cubic b-spline collocation method for solving the nonlinear klein-gordon equation,

N. Malik, A. Singh, and S. Dahiya, “An improvised cubic b-spline collocation method for solving the nonlinear klein-gordon equation,” Available at SSRN 5123194, 2025

2025
[25]

Estimating production functions through additive models based on regression splines,

V. J. España, J. Aparicio, X. Barber, and M. Esteve, “Estimating production functions through additive models based on regression splines,”European Journal of Operational Research, vol. 312, no. 2, pp. 684–699, 2024

2024
[26]

Electric vehicle charging station data,

C. of Boulder Open Data, “Electric vehicle charging station data,”
[2023]

Available: https://open-data.bouldercolorado.gov/ datasets/95992b3938be4622b07f0b05eba95d4c_0/explore Z

[Online]. Available: https://open-data.bouldercolorado.gov/ datasets/95992b3938be4622b07f0b05eba95d4c_0/explore Z. Zhou et al.:Preprint submitted to Elsevier Page 10 of 10

Pith tools

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

[1] [1]

Global EV outlook 2024,

“Global EV outlook 2024,” International Energy Agency (IEA), Tech. Rep., 2024, Available: https://www.iea.org/reports/global-ev- outlook-2024

work page 2024

[2] [2]

Electric vehicle charging load forecasting: A comparative studyofdeeplearningapproaches,

J.Zhu,Z.Yang,M.Mourshed,Y.Guo,Y.Zhou,Y.Chang,Y.Wei,and S. Feng, “Electric vehicle charging load forecasting: A comparative studyofdeeplearningapproaches,” Energies,vol.12,no.14,p.2692, 2019

work page 2019

[3] [3]

Elec- tric vehicle charging load forecasting considering weather impact,

W. Wang, A. Tang, F. Wei, H. Yang, L. Xinran, and J. Peng, “Elec- tric vehicle charging load forecasting considering weather impact,” Applied Energy, vol. 383, p. 125337, 2025

work page 2025

[4] [4]

A transfer learning method for electric vehicles charging strategy based ondeepreinforcementlearning,

K. Wang, H. Wang, Z. Yang, J. Feng, Y. Li, J. Yang, and Z. Chen, “A transfer learning method for electric vehicles charging strategy based ondeepreinforcementlearning,” AppliedEnergy,vol.343,p.121186, 2023

work page 2023

[5] [5]

Review on schedul- ing,clustering,andforecastingstrategiesforcontrollingelectricvehi- cle charging: Challenges and recommendations,

A. S. Al-Ogaili, T. J. T. Hashim, N. A. Rahmat, A. K. Ramasamy, M. B. Marsadek, M. Faisal, and M. A. Hannan, “Review on schedul- ing,clustering,andforecastingstrategiesforcontrollingelectricvehi- cle charging: Challenges and recommendations,”Ieee Access, vol. 7, pp. 128353–128371, 2019

work page 2019

[6] [6]

R. H. Shumway, D. S. Stoffer, and D. S. Stoffer,Time series analysis and its applications. Springer, 2000, vol. 3

work page 2000

[7] [7]

Informer:Beyondefficienttransformerforlongsequencetime-series forecasting,

H.Zhou,S.Zhang,J.Peng,S.Zhang,J.Li,H.Xiong,andW.Zhang, “Informer:Beyondefficienttransformerforlongsequencetime-series forecasting,” in Proceedings of the AAAI conference on artificial intelligence, vol. 35, no. 12, 2021, pp. 11106–11115

work page 2021

[8] [8]

Atimeseries is worth 64 words: Long-term forecasting with transformers,

Y.Nie,N.H.Nguyen,P.Sinthong,andJ.Kalagnanam,“Atimeseries is worth 64 words: Long-term forecasting with transformers,”arXiv preprint arXiv:2211.14730, 2022

arXiv 2022