Pith. sign in

REVIEW 5 major objections 7 minor 32 references

e-Fold Cross-Validation for Recommender-System Evaluation

T0 review · 5 major / 7 minor · reviewed 2026-08-12 · deepseek-v4-flash

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

Pith's one-line read e-fold cross-validation, which stops folding once the running score's confidence interval stabilizes, used 41.5% of the energy of 10-fold cross-validation while producing scores that differed by 1.81% on average.

desk verdict A clean domain-extension study of the authors' own e-CV idea, but the headline energy and reliability numbers are not yet supported: energy is a fold-count proxy, alpha is unreported, and there is no control for random stopping. read the letter →

arxiv 2412.01011 v1 pith:VBXSXSJT submitted 2024-12-02 cs.LG cs.IR

Moritz Baumgart , Lukas Wegmeth , Tobias Vente , Joeran Beel This is my paper · ORCID
classification cs.LGcs.IR
keywords e-foldcross-validationenergyefficiencyrecommendersystemsk-foldearlystoppingconfidenceintervalgreenNDCG@10
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 proposes e-fold cross-validation (e-CV), an early-stopping version of k-fold cross-validation for evaluating recommender systems. The idea is to run folds one at a time and stop as soon as the confidence interval around the running mean score stabilizes, so that the number of folds is chosen by the data rather than fixed in advance. In experiments with five recommender algorithms and six datasets, e-CV stopped after 4.15 folds on average, used 41.5% of the energy of a full 10-fold run, and produced scores that differed from 10-fold results by 1.81% on average. The authors conclude that e-CV is a useful energy-saving alternative that keeps evaluation results and algorithm rankings largely intact.

What carries the argument

The mechanism is the e-CV stopping rule. After each fold, e-CV computes the cumulative mean and the confidence-interval width $c_n$ of the scores seen so far, and stops when $|c_{n-1} - c_n| \leq \alpha/c_n$, where $\alpha$ is a user-selected parameter that trades energy saving against accuracy. The intuition is that once the width stops changing, additional folds are unlikely to move the mean much. The evaluation feeds precomputed 10-fold scores in 5000 different fold orders and records the stopping point and final score for each order.

What would settle it

A direct test would be to run e-CV on a dataset/algorithm pair where the per-fold scores are non-stationary so that the confidence-interval width stabilizes early while later folds shift the mean; if on many permutations the final e-CV score differs from the 10-fold score by substantially more than 1.81%, the claimed reliability does not hold.

Watch

Extended reading notes

Core claim

The central discovery is that a simple criterion on the width of the confidence interval can reliably decide when to stop folding. The paper simulates e-CV by feeding it precomputed scores from a fixed 10-fold split in 5000 random fold orderings, and finds that the stopping rule typically fires around fold 4, with a final NDCG@10 score only 1.81% away from the full 10-fold score. Algorithm rankings produced by e-CV nearly always match those produced by 10-CV. The paper takes this as evidence that e-CV is a viable energy-efficient alternative to standard 10-fold cross-validation.

Load-bearing premise

The load-bearing premise is that a stabilizing confidence-interval width signals that the cumulative mean is close to the full 10-fold mean; the stopping criterion is an ad hoc heuristic with no formal error bound.

Editorial extensions

If this is right

  • If e-CV works as reported, researchers can cut the energy cost of recommender-system evaluation by more than half while staying within about two percentage points of the standard 10-fold result.
  • It would give practitioners a data-dependent choice of fold count instead of the conventional $k=10$, and the ranking of algorithms would remain stable on typical datasets.
  • The energy saving would grow with dataset and model size, because the avoided folds are the expensive training runs.
  • The observed variability across datasets suggests that the method's reliability should be checked per dataset before use.

Reading between the lines

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

  • The stopping rule is a heuristic, so a formal bound on the gap between e-CV and full k-CV would be needed before relying on it for leaderboard-critical comparisons; the paper does not supply one.
  • Because the experiments simulate e-CV from precomputed 10-fold scores, the reported 41.5% energy figure assumes savings scale linearly with fold count; a real deployment should measure end-to-end energy including confidence-interval computation.
  • An adversarial fold ordering, for example scores sorted so that the interval appears stable early, could break the rule, so worst-case performance remains an open question.
  • The same criterion could be tested on other evaluation metrics and other supervised-learning tasks, not only NDCG@10 in recommender systems.
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

5 major / 7 minor

Summary. The paper proposes e-fold cross-validation (e-CV), an early-stopping alternative to k-fold cross-validation for recommender-system evaluation. The method computes the cumulative mean of per-fold NDCG@10 scores and a confidence interval for that mean, then stops folding once the change in confidence-interval width satisfies |c_{n-1} - c_n| <= alpha / c_n. The authors simulate e-CV by reordering the already-computed 10-fold CV scores of 5 algorithms on 6 datasets over 5000 random permutations per dataset-algorithm pair, compare the early-stopped mean to the full 10-fold mean, and report an average percentage difference of 1.81% at an average stopping point of 4.15 folds, which they translate into 41.5% of the energy of 10-CV. They also report that the average ranking of algorithms is largely preserved. The paper concludes that e-CV is a promising energy-efficient and reliable alternative to k-fold cross-validation.

Significance. If substantiated, the claim that a simple stopping rule can reduce cross-validation cost by more than half while changing results by under 2% would be practically valuable for energy-conscious recommender-system research, and the topic is timely given growing attention to green machine learning. The paper uses a reasonable spread of algorithms and datasets, and the permutation-based simulation is a sensible way to explore fold-order dependence. However, the current evidence is not sufficient to support the central claim: the stopping rule has no demonstrated connection between confidence-interval stabilization and accuracy of the early-stopped mean; the free parameter alpha is never reported; the energy claim is based on fold count rather than measured energy; and the simulation uses 10-CV scores rather than true e-fold scores, so it evaluates an early-stopped 10-CV rather than a genuine e-fold procedure. The paper also lacks error bars, a random-stopping baseline, and any statistical test for the reported averages. These are fixable in a revision, but they are load-bearing for the paper's main conclusion.

major comments (5)
  1. [Section 2, stopping rule] The stopping rule |c_{n-1} - c_n| <= alpha / c_n tests only that the confidence-interval width is changing slowly; it does not test whether the cumulative mean is close to the eventual 10-fold mean, and no derivation, error bound, or calibration study connects these two quantities. Moreover, the value of alpha is never reported, so the headline numbers (4.15 folds, 1.81% difference) are not reproducible and could reflect a favorable parameter choice. The authors should report alpha, provide a sensitivity analysis over alpha, and ideally replace or supplement the heuristic with a rule that has a formal or empirical convergence guarantee.
  2. [Section 2, simulation methodology] The evaluation is a simulation of early stopping within a fixed 10-fold split, not of an e-fold cross-validation procedure. Each of the 10 scores used in the simulation comes from a model trained on 90% of the data, whereas a true e-fold run with e around 4 would train each model on only 75% of the data, changing both the scores and the stopping behavior. The reported 41.5% energy saving is therefore an estimate for early-stopped 10-CV and cannot be directly attributed to e-CV without retraining models with the selected number of folds (or at least a careful discussion of why the training-set-size mismatch is negligible).
  3. [Section 3, energy claim] The statement that e-CV 'only needed 41.5% of the energy that 10-fold cross validation would need' equates stopping after an average of 4.15 folds with using 41.5% of the energy, but no energy is measured. Energy per fold is not constant: it depends on model type, dataset size, training-set size, and hardware, and for the deep models considered the cost of training on 90% versus 75% of the data is not generally proportional to the number of folds. At most the paper can claim an average of 4.15 test folds, or about 41.5% of the number of model trainings in a 10-CV run, not 41.5% of energy.
  4. [Figure 2 and Section 3, averaging] Several dataset-algorithm cells in Figure 2 show percentage differences well above the 1.81% average, especially for MultiVAE and Pop, with some bars reaching roughly 5-6%. The 1.81% average across 30 cells and 5000 permutations hides a heavy upper tail, and no standard deviations, quantiles, or error bars are reported despite the 5000 permutations. The authors should report the distribution of percentage differences (e.g., median, 90th percentile, worst-case cell) rather than only the mean, to allow readers to assess reliability in the worst-performing settings.
  5. [Section 3, no comparison baseline] There is no baseline comparison, such as a random stopping rule that halts at the same average fold count, and no statistical test comparing e-CV's 1.81% difference to that baseline. Without such a comparison, the CI-based stopping criterion cannot be distinguished from simply deciding to stop early after a fixed number of folds, and the claim that the criterion adds reliability beyond early stopping in general is not established.
minor comments (7)
  1. [Abstract] The phrase "it's results" should be "its results."
  2. [Section 1] "Our intension was" should be "Our intention was."
  3. [Section 2] The confidence interval is never formally defined; the authors should give the exact formula for c_n (e.g., standard error times a critical value, or a bootstrap interval) and state what distributional assumption is used.
  4. [Table 1] The table uses commas as decimal separators inconsistently (e.g., "0,0669" vs. "7,8049") and the Density column lacks units or a clarifying caption; use a consistent decimal notation.
  5. [Figures 2 and 3] The bars in Figures 2 and 3 would be easier to read if the average values were labeled, and since they are averages over 5000 permutations, the figures should include error bars or at least a stated measure of dispersion.
  6. [References] Reference [6] (Bergman et al., 'Don't waste your time: Early stopping cross-validation') is directly related and should be compared explicitly; the current text only cites it in passing, so the reader cannot see how e-CV differs from that prior early-stopping method.
  7. [Figure 4] The claim that rankings 'stayed consistent' would be strengthened by reporting a rank-correlation coefficient (e.g., Kendall's tau) and the fraction of permutations in which the algorithm ranking differs between e-CV and 10-CV.

Circularity Check

0 steps flagged · score 1.0 of 10

No significant circularity: e-CV is benchmarked against full 10-fold CV on the same folds, and the claimed energy/accuracy trade-off is an empirical measurement rather than a consequence of the method's definition.

full rationale

The paper's central claim is that e-CV stops after 4.15 folds on average and differs from 10-fold CV by 1.81%. This is an empirical comparison: e-CV's cumulative mean at the stopping fold is compared with the full 10-fold mean, and the energy ratio follows from the fold count. Nothing in the definition of the stopping rule (|c_{n-1} - c_n| ≤ α/c_n) forces the reported 1.81% or 41.5%; those numbers depend on the datasets, algorithms, permutations, and the unreported value of α. The self-citations [5], [21], and [27] position the work and motivate the energy concern, but they do not carry the empirical claim. The stopping rule is a heuristic whose validity is not derived, and α is not reported, which are reproducibility and validity concerns, not circularity. Because the self-references are contextual and non-load-bearing, the score is 1 rather than 0.

Assumptions & free parameters 1 free parameters · 4 assumptions · 0 invented entities

The central claim rests on an unreported stopping threshold (alpha), a linear energy-scaling assumption that equates fold count with energy, and an unstated confidence-interval formula. The permutation sampling and the choice of 10-CV as ground truth are additional assumptions. No new entities are introduced.

free parameters (1)
  • alpha (stopping threshold) = Not reported
    User-selectable parameter in the stopping criterion; its value directly controls the trade-off between energy savings and accuracy. The paper never states the value used in the experiments, so the reported 1.81% and 4.15 folds are conditional on an unspecified alpha.
assumptions (4)
  • domain assumption Energy consumption scales linearly with the number of folds used
    The paper equates energy usage to the fraction of folds run (41.5% energy for 4.15 folds), but never measures energy. This ignores fixed overheads and makes the energy claim a proxy rather than a measurement.
  • standard math The confidence interval width is computed with a specific, unstated formula (e.g., normal or t-distribution)
    The paper says it calculates the confidence interval of the mean but does not specify the formula, the confidence level, or how the CI width is defined. This affects the stopping criterion.
  • domain assumption The 5000 random permutations are representative of the 10! possible fold orders
    Only 5000 of 3,628,800 permutations are sampled, with no seed or variance estimate, so the averages over permutations may not be stable.
  • domain assumption 10-CV is an appropriate ground truth
    The paper compares e-CV to 10-CV as ground truth, citing prior work that 5-10 folds is optimal. This is a reasonable domain assumption but not justified in this paper.

how reviews work

0 comments
Cite this review

Pith. "Pith review of e-Fold Cross-Validation for Recommender-System Evaluation." pith.science (2026). https://pith.science/paper/VBXSXSJT

@misc{pith2026241201011,
  author       = {Pith},
  title        = {Pith review of: e-Fold Cross-Validation for Recommender-System Evaluation},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/VBXSXSJT}},
  note         = {Machine review of arXiv:2412.01011}
}
read the original abstract

To combat the rising energy consumption of recommender systems we implement a novel alternative for k-fold cross validation. This alternative, named e-fold cross validation, aims to minimize the number of folds to achieve a reduction in power usage while keeping the reliability and robustness of the test results high. We tested our method on 5 recommender system algorithms across 6 datasets and compared it with 10-fold cross validation. On average e-fold cross validation only needed 41.5% of the energy that 10-fold cross validation would need, while it's results only differed by 1.81%. We conclude that e-fold cross validation is a promising approach that has the potential to be an energy efficient but still reliable alternative to k-fold cross validation.

Figures

Figures reproduced from arXiv: 2412.01011 by the authors.

Figure 1
Figure 1. Exemplary e-CV run Our proposed implementation calculates the mean of all scores it has so far, as well as the confidence interval (CI) of that mean. It then uses a criterion on the CI width to stop folding. Let C = {c1, ..., cn} be the set of CI widths then we stop folding if |cn−1 − cn| ≤ α cn , with α being a user-selectable parameter which can be used to prioritize energy-saving (large α) or accuracy (small α). … view at source ↗
Figure 2
Figure 2. Percentage difference between final e-CV score and 10-CV score for each dataset/algorithm, averaged across tested permutations. 0 1 2 3 4 5 6 7 NeuMF Pop ImplicitMF ItemKNN Mul �VAE NeuMF Pop ItemKNN ImplicitMF Mul �VAE Mul �VAE ImplicitMF NeuMF ItemKNN Pop NeuMF Pop Mul �VAE ImplicitMF ItemKNN ImplicitMF Pop ItemKNN NeuMF Mul �VAE Pop ImplicitMF ItemKNN Mul �VAE NeuMF Amazon2014-Amazon￾Instant-Video Amazon2014-Apps… view at source ↗
Figure 3
Figure 3. Stopping point determined by e-CV for each dataset/algorithm, averaged across tested permutations. Looking at fig. 3, we can see at which fold e-CV decided to stop the folding process. It is notable that for the MovieLens datasets it stopped at later folds, which could explain the smaller percentage difference for these datasets in fig. 2. This, however, does not hold for the LastFM dataset, where both percentage di… view at source ↗
Figures from the paper (1 more)
Figure 4
Figure 4. Figure 4: Ranking of the algorithms, averaged across tested permutations. [PITH_FULL_IMAGE:figures/full_fig_p005_4.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

32 extracted references · 14 canonical work pages

  1. [5]

    https: //doi.org/10.31219/osf.io/exw3j

    Beel, J., Wegmeth, L., Vente, T.: e-fold cross-validation: A computing and energy- efficient alternative to k-fold cross-validation with adaptive folds (Jun 2024). https: //doi.org/10.31219/osf.io/exw3j

    work page doi:10.31219/osf.io/exw3j 2024

  2. [21]

    Mahlich, C., Vente, T., Beel, J.: From theory to practice: Implementing and eval- uating e-fold cross-validation. In: International Conference on Artificial Intelli- gence and Machine Learning Research (CAIMLR) (2024), https://isg.beel.org/ blog/2024/09/16/e-fold-cross-validation/ e-Fold Cross-Validation for Recommender-System Evaluation 7

    work page 2024

  3. [28]

    Wegmeth, L., Vente, T., Said, A., Beel, J.: Emers: Energy meter for recommender systems. In: International Workshop on Recommender Systems for Sustainability and Social Good (RecSoGood) at the 18th ACM Conference on Recommender Systems (ACM RecSys) (2024), https://arxiv.org/pdf/2409.15060

    work page Pith review arXiv 2024

  4. [6]

    Bergman, E., Purucker, L., Hutter, F.: Don’t waste your time: Early stopping cross-validation (2024), https://arxiv.org/abs/2405.03389

    work page Pith review arXiv 2024

  5. [1]

    Anguita, D., Ghelardoni, L., Ghio, A., Oneto, L., Ridella, S.: The ’k’ in k-fold cross validation. In: 20th European Symposium on Artificial Neural Networks, ESANN 2012, Bruges, Belgium, April 25-27, 2012 (2012), https://www.esann.org/sites/ default/files/proceedings/legacy/es2012-62.pdf

    work page 2012

  6. [2]

    Arabzadeh, A., Vente, T., Beel, J.: Green recommender systems: Optimizing dataset size for energy-efficient algorithm performance. In: International Work- shop on Recommender Systems for Sustainability and Social Good (RecSoGood) at the 18th ACM Conference on Recommender Systems (ACM RecSys) (2024), https://isg.beel.org/pubs/2024-Green-RecSys-Dataset-Samp...

    work page 2024

  7. [3]

    Statistics Surveys 4(none), 1–50 (2016)

    Arlot, S., Celisse, A.: A survey of cross-validation procedures for model selection. Statistics Surveys 4(none), 1–50 (2016). https://doi.org/10.1214/09-ss054

    doi:10.1214/09-ss054 2016

  8. [4]

    Recommender-Systems.com Blog (2024)

    Beel, J., Said, A., Vente, T., Wegmeth, L.: Green recommender sys- tems – a call for attention. Recommender-Systems.com Blog (2024). https://doi.org/10.31219/osf.io/5ru2g, https://isg.beel.org/pubs/2024_Green_ Recommender_Systems-A_Call_for_Attention.pdf

    doi:10.31219/osf.io/5ru2g 2024

Show all 32 references
  1. [7]

    In: Proceedings of the fifth ACM conference on Recommender systems

    Cantador, I., Brusilovsky, P., Kuflik, T.: Second workshop on information hetero- geneity and fusion in recommender systems (hetrec2011). In: Proceedings of the fifth ACM conference on Recommender systems. RecSys ’11, ACM (Oct 2011). https://doi.org/10.1145/2043932.2044016 6 B...

    2011

  2. [9]

    https://doi.org/10.7930/j0r49nqx

    Crimmins, A., Balbus, J., Gamble, J., Beard, C., Bell, J., Dodgen, D., Eisen, R., Fann, N., Hawkins, M., Herring, S., Jantarasami, L., Mills, D., Saha, S., Sarofim, M., Trtanj, J., Ziska, L.: The Impacts of Climate Change on Human Health in the United States: A Scientific Asse...

    2016 doi

  3. [10]

    ACM Transactions on Information Systems22(1), 143–177 (Jan 2004)

    Deshpande, M., Karypis, G.: Item-based top-nrecommendation algorithms. ACM Transactions on Information Systems22(1), 143–177 (Jan 2004). https://doi.org/ 10.1145/963770.963776

    2004

  4. [11]

    In: Proceedings of the 29th ACM International Conference on Information & Knowledge Management

    Ekstrand, M.D.: Lenskit for python: Next-generation software for recommender systems experiments. In: Proceedings of the 29th ACM International Conference on Information & Knowledge Management. CIKM ’20, ACM (Oct 2020). https: //doi.org/10.1145/3340531.3412778

    2020

  5. [12]

    In: 2021 IEEE International Symposium on High-Performance Computer Architecture (HPCA)

    Gupta, U., Kim, Y.G., Lee, S., Tse, J., Lee, H.H.S., Wei, G.Y., Brooks, D., Wu, C.J.: Chasing carbon: The elusive environmental footprint of computing. In: 2021 IEEE International Symposium on High-Performance Computer Architecture (HPCA). IEEE (Feb 2021). https://doi.org/10.1...

    2021

  6. [13]

    ACM Transactions on Interactive Intelligent Systems5(4), 1–19 (Dec 2015)

    Harper, F.M., Konstan, J.A.: The movielens datasets: History and context. ACM Transactions on Interactive Intelligent Systems5(4), 1–19 (Dec 2015). https://doi. org/10.1145/2827872

    2015 doi

  7. [14]

    https://doi.org/10.1145/ 2872427.2883037

    He, R., McAuley, J.: Ups and downs: Modeling the visual evolution of fashion trends with one-class collaborative filtering (Apr 2016). https://doi.org/10.1145/ 2872427.2883037

    2016

  8. [15]

    In: Proceedings of the 26th International Conference on World Wide Web

    He, X., Liao, L., Zhang, H., Nie, L., Hu, X., Chua, T.S.: Neural collaborative filtering. In: Proceedings of the 26th International Conference on World Wide Web. WWW ’17, International World Wide Web Conferences Steering Committee (Apr 2017). https://doi.org/10.1145/3038912.3052569

    2017

  9. [16]

    https://doi.org/10.48550/ARXIV.1808.09781

    Kang, W.C., McAuley, J.: Self-attentive sequential recommendation (Aug 2018). https://doi.org/10.48550/ARXIV.1808.09781

    2018 doi arXiv

  10. [17]

    Kohavi, R., John, G.: The Wrapper Approach, vol. 14, pp. 33–50. Springer US (2001). https://doi.org/10.1007/978-1-4615-5725-8_3, https: //www.researchgate.net/publication/2352264_A_Study_of_Cross-Validation_ and_Bootstrap_for_Accuracy_Estimation_and_Model_Selection

    2001

  11. [19]

    In: Proceedings of the 2018 World Wide Web Conference on World Wide Web - WWW ’18

    Liang, D., Krishnan, R.G., Hoffman, M.D., Jebara, T.: Variational autoencoders for collaborative filtering. In: Proceedings of the 2018 World Wide Web Conference on World Wide Web - WWW ’18. pp. 689–698. WWW ’18, ACM Press (2018). https://doi.org/10.1145/3178876.3186150

    2018

  12. [20]

    In: Proceedings of the 17th ACM Conference on Recommender Systems

    Ma, H., Xie, R., Meng, L., Chen, X., Zhang, X., Lin, L., Zhou, J.: Exploring false hard negative sample in cross-domain recommendation. In: Proceedings of the 17th ACM Conference on Recommender Systems. RecSys ’23, ACM (Sep 2023). https://doi.org/10.1145/3604915.3608791

    2023

  13. [22]

    Marcot, B.G., Hanea, A.M.: What is an optimal value of k in k-fold cross-validation in discrete bayesian network analysis? Comput. Stat. 36(3), 2009–2031 (2021). https://doi.org/10.1007/S00180-020-00999-9

    2021 doi

  14. [23]

    IEEE Access 6, 69009–69022 (2018)

    Mu, R.: A survey of recommender systems based on deep learning. IEEE Access 6, 69009–69022 (2018). https://doi.org/10.1109/access.2018.2880197

    2018

  15. [24]

    Refaeilzadeh, P., Tang, L., Liu, H.: Cross-Validation, pp. 1–7. Springer New York (2009). https://doi.org/10.1007/978-1-4899-7993-3_565-2

    2009 doi

  16. [25]

    In: Proceedings of the 28th ACM International Conference on Information and Knowledge Management

    Sun, F., Liu, J., Wu, J., Pei, C., Lin, X., Ou, W., Jiang, P.: Bert4rec: Sequen- tial recommendation with bidirectional encoder representations from transformer. In: Proceedings of the 28th ACM International Conference on Information and Knowledge Management. CIKM ’19, ACM (No...

    2019

  17. [26]

    arXiv / Journal of Artificial Intelligence Research77, 427–457 (2021 / 2023)

    Tornede, T., Tornede, A., Hanselle, J., Mohr, F., Wever, M., Hüllermeier, E.: To- wards green automated machine learning: Status quo and future directions. arXiv / Journal of Artificial Intelligence Research77, 427–457 (2021 / 2023)

    2021

  18. [27]

    In: Proceedings of the 18th ACM Conference on Recommender Systems

    Vente, T., Wegmeth, L., Said, A., Beel, J.: From clicks to carbon: The environmen- tal toll of recommender systems. In: Proceedings of the 18th ACM Conference on Recommender Systems. p. 580–590. RecSys ’24, Association for Computing Ma- chinery, New York, NY, USA (2024). https...

    2024

  19. [29]

    In: 2016 IEEE 6th Interna- tional Conference on Advanced Computing (IACC)

    Yadav, S., Shukla, S.: Analysis of k-fold cross-validation over hold-out valida- tion on colossal datasets for quality classification. In: 2016 IEEE 6th Interna- tional Conference on Advanced Computing (IACC). IEEE (Feb 2016). https: //doi.org/10.1109/iacc.2016.25

    2016 doi

  20. [30]

    In: 2017 51st Asilomar Conference on Signals, Systems, and Computers

    Yang, T.J., Chen, Y.H., Emer, J., Sze, V.: A method to estimate the energy con- sumption of deep neural networks. In: 2017 51st Asilomar Conference on Signals, Systems, and Computers. IEEE (Oct 2017). https://doi.org/10.1109/acssc.2017. 8335698

    2017 doi

  21. [31]

    ACM Computing Surveys52(1), 1–38 (Feb 2019)

    Zhang, S., Yao, L., Sun, A., Tay, Y.: Deep learning based recommender system: A survey and new perspectives. ACM Computing Surveys52(1), 1–38 (Feb 2019). https://doi.org/10.1145/3285029

    2019 doi

  22. [32]

    Journal of Econometrics187(1), 95–112 (7 2015)

    Zhang, Y., Yang, Y.: Cross-validation for selecting a model selection proce- dure. Journal of Econometrics187(1), 95–112 (7 2015). https://doi.org/10.1016/ j.jeconom.2015.02.006

    2015

  23. [33]

    https://doi.org/10.48550/ARXIV.2011

    Zhao, W.X., Mu, S., Hou, Y., Lin, Z., Chen, Y., Pan, X., Li, K., Lu, Y., Wang, H., Tian, C., Min, Y., Feng, Z., Fan, X., Chen, X., Wang, P., Ji, W., Li, Y., Wang, X., Wen, J.R.: Recbole: Towards a unified, comprehensive and efficient framework for recommendation algorithms (No...

    2020 doi

  24. [34]

    In: Proceedings of the 17th ACM Conference on Recommender Systems

    Zhu, J., Wang, Y., Zhu, F., Sun, Z.: Domain disentanglement with interpolative data augmentation for dual-target cross-domain recommendation. In: Proceedings of the 17th ACM Conference on Recommender Systems. RecSys ’23, ACM (Sep 2023). https://doi.org/10.1145/3604915.3608802

    2023

Pith tools

Reviewed August 12, 2026 · model on record in the stance chip above.