Pith. sign in

REVIEW 4 major objections 5 minor 58 references

Industrial-scale Prediction of Cement Clinker Phases using Machine Learning

T0 review · 4 major / 5 minor · reviewed 2026-08-11 · deepseek-v4-flash

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

Pith's one-line read Trained on two years of data from a working cement plant, a machine-learning pipeline predicts the three main clinker mineral phases from process readings, beating the industry-standard Bogue equations used for quality control.

desk verdict Useful industrial dataset and a fair Bogue comparison, but the headline accuracy claims are inflated by post-production inputs and a mis-specified time alignment; needs revision before the real-time claim holds. read the letter →

arxiv 2412.11981 v2 pith:74GGEPXL submitted 2024-12-16 cs.LG cond-mat.mtrl-sci

Sheikh Junaid Fayaz , Nestor Montiel-Bohorquez , Shashank Bishnoi , Matteo Romano , Manuele Gatti , N. M. Anoop Krishnan This is my paper · ORCID
classification cs.LGcond-mat.mtrl-sci
keywords machinelearningcementclinkerphasesBogueequationalitebeliteferritedigitaltwinprocessparameters
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 claims that data-driven models trained on two years of operation at an industrial cement plant can predict the three main clinker mineral phases — alite, belite, and ferrite — accurately enough for real-time quality control. With all 59 plant measurements as input, a neural network predicts alite with a mean absolute percentage error of 1.24%, a Gaussian-process model predicts belite at 6.77%, and a support-vector model predicts ferrite at 2.53%; restricting inputs to process parameters alone still beats the Bogue equations, for example alite at 3.14% error versus Bogue's 7.79%. The paper also fits plant-specific linear “clinker equations” from the same data that keep Bogue's pen-and-paper simplicity while roughly halving its error. If these results hold, a cement plant could estimate clinker quality minutes before the clinker is made, reduce off-specification waste, and adjust the process proactively instead of after the fact.

What carries the argument

The machinery is the synchronized, cleaned plant archive combined with per-phase model selection. The paper's temporal alignment protocol assigns each clinker composition to process conditions roughly 37 minutes earlier (a 1-minute buffer, 16-minute preheater residence, 20-minute cooler retention, and 20-minute sampling delay) and averages all signals to two-hour windows; after removing duplicates, missing entries, negative values, and values outside the 0.01–99.99 percentile range, 8,654 rows remain. Eight regression architectures are compared on fifteen input-feature combinations, and the best per-phase models are chosen by cross-validation. The Bogue equation — the classical linear recipe that derives clinker phase percentages from oxide content — serves as the baseline to beat; the paper also extracts plant-specific linear equations from the same data as an intermediate tool between Bogue and full black-box models.

What would settle it

Hold out the entire second year (or any continuous 60-day block), train on the remaining 18 months, and compute the alite MAPE on that held-out block; if it climbs well above 1.24% and toward the 3–4% range, the published accuracy reflects split leakage rather than forecasting skill. A second check: rerun the pipeline with the 19.4% of rows removed as “outliers” restored and see whether the model still tracks the rapid alite swings it is credited with, or whether those transients were excluded precisely so the model would not have to predict them.

Watch

Extended reading notes

Core claim

The central discovery the authors claim is that the mineral composition of industrial clinker is learnable from operational records with enough accuracy to replace the Bogue calculation in practice. They assembled a two-year dataset of 8,654 complete measurements after synchronizing process readings to clinker samples through a fixed residence-time chain of about 37 minutes (plus sampling delay), and found that non-linear non-parametric models — a neural network for alite, a Gaussian-process model for belite, and support vector regression for ferrite — capture compositional swings that the linear Bogue equations systematically miss. On a two-month hold-out period the models reduced prediction error several-fold relative to Bogue, and the improvement persisted even when the inputs were limited to process parameters that are available before the clinker is produced. A post-hoc Shapley-value analysis shows the learned drivers match established clinker chemistry — CaO and SiO$_2$ dominate alite, Fe$_2$O$_3$ governs ferrite — which the authors read as evidence the models learned real physical relationships rather than noise.

Load-bearing premise

The load-bearing premise is that a fixed time delay of about 37 minutes (plus sampling lag) pairs each clinker measurement with its true process inputs, and that a random 70/30 split of the resulting auto-correlated two-year series yields independent training and test samples.

Editorial extensions

If this is right

  • Real-time control: a model fed only process parameters and kiln-feed composition can estimate clinker quality before the clinker exists, so operators could adjust the kiln instead of discarding out-of-specification product.
  • Cheap plant-specific equations: a plant can fit its own linear “clinker equation” from routine XRF/XRD records and immediately beat generic Bogue calculations, with no machine-learning infrastructure required.
  • Transferable digital twin: because the framework is trained purely on plant records, the same pipeline can be retrained for other plants and extended toward properties such as 28-day strength and alternative fuels and raw materials.
  • Interpretable chemistry: the learned feature attributions give quantitative, plant-specific rankings of how each oxide drives phase formation, complementing textbook clinker chemistry with measured local relationships.

Reading between the lines

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

  • The stated accuracy likely depends on split physics more than the paper reports: clinker is sampled hourly and the series is strongly autocorrelated, so neighboring records land on opposite sides of the random 70/30 split and shadow each other; a time-ordered test (train on the first year, predict the second) would reveal how much of the 1.24% alite error is true forecasting.
  • The 19.4% of rows dropped as percentile “outliers” are precisely the plant upsets the model is credited with tracking, including the rapid one-day alite swings; retraining on the unfiltered record would show whether the model learned to predict transients or simply had them removed.
  • If the process-parameters-only result survives a blocked temporal test, it implies that kiln temperature, pressure, fuel, and airflow readings encode most of the phase-formation information, which would guide other plants toward sensor-first data collection without expensive online XRD.
  • Because the plant-specific linear equations fit from routine XRF/XRD records already beat Bogue, a short campaign of paired measurements at any plant could yield a local, low-cost quality-control formula even before a full machine-learning digital twin is built.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

4 major / 5 minor

Summary. The paper develops machine-learning models to predict clinker phase compositions (alite, belite, ferrite) from two years of operational data from an industrial cement plant. Fifteen feature combinations built from process parameters (PP), kiln feed (KF), hot meal (HM), and clinker oxides (CO) are tested across nine regression architectures, with performance measured by MAPE, MAE, and R2. The authors report that the best full-feature models achieve MAPEs of 1.24% (alite), 6.77% (belite), and 2.53% (ferrite), that PP-only models still outperform the plant-specific Bogue equation (e.g., alite MAPE 3.14% vs 7.79%), and that data-driven plant-specific linear equations provide a middle ground between Bogue and black-box ML. SHAP analysis is used to attribute predictions to clinker oxides. The central claim is that ML-based digital twins can enable real-time clinker quality control and outperform conventional Bogue calculations.

Significance. If the reported results are valid, the paper would offer a practically valuable demonstration that data-driven models can predict clinker mineralogy from process data in a real industrial setting, with clear implications for quality control and emissions reduction. Strengths include the unusually large industrial dataset (two years, ~8,654 curated measurements), a systematic comparison across many model architectures and feature sets, a direct benchmark against the industry-standard Bogue equation, and public release of the code. The SHAP analysis provides a useful interpretability layer. However, the significance is currently limited by several evaluation issues that affect the trustworthiness of the headline numbers: the temporal alignment of inputs and outputs omits the rotary kiln residence time; the train/test split is random on an autocorrelated series and is described inconsistently; and the best accuracy figures come from models that include clinker oxides measured on the same sample as the XRD phase labels, which is a post-production mapping rather than a forward prediction. These issues need to be addressed before the claimed 'unprecedented accuracy' can be accepted.

major comments (4)
  1. [§2.1 and §3.1] The synchronization timeline is internally inconsistent. The stated sequence is 1-minute buffer, 16-minute preheater residence, and 20-minute cooler retention, totaling ~37 minutes, but the rotary kiln—where clinker phases actually form and which has a typical material residence time of 20–30 minutes—is absent from the timeline. As a result, process and feed inputs are aligned to clinker outputs roughly one kiln dwell time too early. Because every training sample depends on this alignment, all reported metrics (including the two-month 'unseen' evaluation in Fig. 2) are affected. The 2-hour averaging may partially mask the error, but it will systematically bias the learned mapping, particularly during the rapid transients the paper emphasizes (e.g., 15 wt.% daily alite swings). The authors must correct the timeline with plant-specific residence-time data, or at minimum perform a sensitivity analysis over plausible kiln residence times and show that the conclusions are unchanged.
  2. [§2.5 and §3.1] The data-splitting protocol is reported inconsistently and is inappropriate for a time series. Section 2.5 states an 80:20 train-test split, while Section 3.1 states a 70:30 split. More importantly, the random split of a two-year autocorrelated series means that temporally adjacent points fall into both training and test sets, inflating apparent accuracy through information leakage. The two-month January–February 2020 evaluation is a good idea, but it does not repair the leakage from the random split used for model selection and hyperparameter tuning. The authors should use a strictly temporal split (e.g., train on an early contiguous block and test on a later contiguous block) and report the autocorrelation of the target series and of the residuals, so that the reader can judge the effective number of independent test samples.
  3. [§3.2.1, Fig. 2, and Table 2] The headline accuracy (alite MAPE 1.24%) comes from models that include clinker oxides (CO) as inputs. Since CO and the XRD phase labels are measured on the same clinker sample, this is a post-production mapping from XRF chemistry to XRD phase composition—similar in information content to the Bogue equation, which also takes clinker oxides as inputs. Framing this as 'predicting' clinker phases 'while requiring minimal inputs' (Abstract) is misleading; the CO-based models are not usable for real-time process control, as the paper itself acknowledges. The fair comparison to Bogue for forward prediction is the PP-only model (alite MAPE 3.14%), which still shows a meaningful improvement. The paper should restructure the claims so that the post-production XRF-to-XRD mode is clearly separated from the forward predictive mode, and should avoid citing the 1.24% figure as the central achievement.
  4. [§2.2 and Table A.5] The 0.01–99.99 percentile filtering removes 2,910 rows (19.4% of the data after earlier cleaning, and about 42% of the raw 14,985 clinker measurements), and it is applied uniformly to all 59 input features and 3 outputs. This approach can systematically exclude the very transients that the model is supposed to predict—the paper highlights 15 wt.% alite swings as a success, but those swings may be underrepresented after filtering. The authors should report model performance without this percentile filter (or with a milder filter) and demonstrate that the removed points are measurement artifacts rather than genuine process variability. At a minimum, the sensitivity of the reported MAPEs to the filtering threshold should be documented.
minor comments (5)
  1. [§2.4, Eq. (3)] The definition of R2 and the auxiliary quantities is garbled. The text defines TSS as the sum of squared differences between predictions and a quantity Äüyý that is itself defined as the mean of squared prediction errors, which is not the mean of the true values. The standard definition should be TSS = Σ(y_t(i) − Ä«y_t)², with Ä«y_t the mean of the true targets.
  2. [Appendix D] The data-driven clinker equations contain typographical errors that make them unusable as printed. In Case 1, the alite equation is written as Alite = 2.97CaO − 4.5SiO2 − 7.25SiO2 + 0.05Fe2O3, with SiO2 appearing twice; the coefficient −7.25 is almost certainly intended for Al2O3. The corresponding matrix also lists two SiO2 columns. Additionally, Cases 2 and 4 report identical intercept vectors (−166.9, −219.4, −45), which is suspicious because the two cases use different feature sets. These equations should be corrected and re-validated.
  3. [§2.3 and Fig. 2 caption] The text in Section 2.3 says 'eight machine learning architectures' and lists linear regression, lasso, elastic net, support vector regression, random forest, XGBoost, neural networks, and Gaussian processes, while Fig. 2 and Table 2 include ridge in addition (nine models total). The count and the model lists should be harmonized.
  4. [§2.6, Eq. (4)] The displayed Shapley value formula is not written correctly: the summation index and the power set notation are confusing, and the expression is missing the necessary subtraction and averaging structure of the standard Shapley formula. Please replace it with a standard, correctly typed version (or refer to the original reference without reproducing a malformed equation).
  5. [Data availability] The underlying dataset is proprietary and not available, while the code is public. This limits reproducibility; at least the preprocessing and synchronization steps should be described with enough detail (including exact timestamps, units, and merging rules) that the pipeline could be re-run on a similar plant dataset.

Circularity Check

0 steps flagged · score 2.0 of 10

No significant circularity: the reported predictions are independent supervised fits, and the self-citations are methodological and non-load-bearing.

full rationale

The derivation chain is not circular under the definitions in the review protocol. Target variables (XRD alite, belite, ferrite) never appear as input features, and no fitted parameter is renamed as a prediction: models are trained on a training split and scored on a held-out test set and on an excluded two-month period, and the plant-specific clinker equations in Appendix D are linear regressions fitted to the training portion and reported with separate train/test R2 values. The CO-input models map XRF oxide compositions measured on the same clinker sample to XRD phase fractions; this is a calibration-style prediction rather than a process-data prediction, and the paper itself states that CO-containing models are post-production and 'inhibit predictive control,' while the PP-only models (alite MAPE 3.14%) independently outperform the Bogue baseline. The SHAP analysis restates known clinker chemistry (CaO positive and SiO2 negative for alite) and is presented as corroboration, not as a derivation. Self-citations [19], [34], and [43] point to the corresponding author's methodological works for standard ML formulas, hyperparameter protocols, and SHAP implementation; they are not load-bearing, no uniqueness theorem is invoked, and no ansatz is smuggled via citation. The fixed ~37-minute alignment in Section 2.1 omits rotary-kiln residence time and the random 70/30 split of an autocorrelated series risks leakage; these are substantive correctness and statistical-validation concerns, but they are not definitional circularity because no equation reduces a claimed prediction to its own input by construction. A minor score of 2 reflects only the presence of non-load-bearing self-citations; the central claim retains independent empirical content.

Assumptions & free parameters 3 free parameters · 4 assumptions · 0 invented entities

The central claims rest on the proprietary measured dataset and standard ML practice. The plant-specific coefficients and outlier thresholds are fitted to the data, and the residence-time constants are assumed. No new physical entities are introduced.

free parameters (3)
  • Plant-specific clinker equation coefficients = Case 1-4 matrices in Appendix D (e.g., alite coefficients 2.97, -4.5, -7.25, 0.05)
    Fit by linear regression to the 8,654-sample dataset; the central 'clinker equations' result depends on these fitted coefficients.
  • Outlier percentile thresholds = 0.01-99.99 percentile per variable
    Chosen by hand; removes 19.4% of rows and shapes the distribution the models are trained and tested on.
  • Residence-time constants = 1, 16, 20, and 20 minutes
    Assumed fixed material transport delays used to synchronize inputs and outputs; no uncertainty or variation reported.
assumptions (4)
  • domain assumption XRD and XRF measurements are accurate ground truth for clinker phases and oxide compositions
    The entire supervised learning setup treats these industrial measurements as labels and features without independent validation (Section 2.1).
  • domain assumption The two-year plant operation is statistically stationary after filtering
    The paper claims the filtered data 'is coming from a single distribution' (Section 3.1), which underpins pooling all data into one train/test split.
  • standard math Standard ML implementations in sklearn are correct
    The paper relies on off-the-shelf algorithms; no custom derivations are checked.
  • standard math SHAP values accurately attribute feature importance for the trained models
    SHAP theory is taken as given (Section 2.6).

how reviews work

0 comments
Cite this review

Pith. "Pith review of Industrial-scale Prediction of Cement Clinker Phases using Machine Learning." pith.science (2026). https://pith.science/paper/74GGEPXL

@misc{pith2026241211981,
  author       = {Pith},
  title        = {Pith review of: Industrial-scale Prediction of Cement Clinker Phases using Machine Learning},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/74GGEPXL}},
  note         = {Machine review of arXiv:2412.11981}
}
read the original abstract

Cement production, exceeding 4.1 billion tonnes and contributing 2.4 tonnes of CO2 annually, faces critical challenges in quality control and process optimization. While traditional process models for cement manufacturing are confined to steady-state conditions with limited predictive capability for mineralogical phases, modern plants operate under dynamic conditions that demand real-time quality assessment. Here, exploiting a comprehensive two-year operational dataset from an industrial cement plant, we present a machine learning framework that accurately predicts clinker mineralogy from process data. Our model achieves unprecedented prediction accuracy for major clinker phases while requiring minimal input parameters, demonstrating robust performance under varying operating conditions. Through post-hoc explainable algorithms, we interpret the hierarchical relationships between clinker oxides and phase formation, providing insights into the functioning of an otherwise black-box model. This digital twin framework can potentially enable real-time optimization of cement production, thereby providing a route toward reducing material waste and ensuring quality while reducing the associated emissions under real plant conditions. Our approach represents a significant advancement in industrial process control, offering a scalable solution for sustainable cement manufacturing.

Figures

Figures reproduced from arXiv: 2412.11981 by the authors.

Figure 1
Figure 1. Dataset characteristics and temporal variability in clinker phases. a, Schematic representation of a cement plant showing key measurement locations: kiln feed (KF), process parameters (PP), hot meal (HM), and clinker oxides (CO). The Venn diagram illustrates the combinations of input features used for model development. b, Two-year temporal evolution of alite content showing plant variability (black dots) with 0.01-… view at source ↗
Figure 2
Figure 2. Performance comparison of machine learning architectures for clinker phase prediction. a-c, Mean Absolute Percentage Error (MAPE) across nine ML models for predicting alite, belite, and ferrite compositions using complete feature sets (KF, PP, HM, CO), respectively. Values in parentheses indicate test set MAPE. The best-performing models are shown in bold. Quantitative performance metrics (R 2 and MAPE) for the best… view at source ↗
Figure 3
Figure 3. Comparison with Bogue equation. a, MAPE of optimal machine learning models (Neural Network for alite, Gaussian Process Regression for belite, Support Vector Regression for ferrite) across 15 combinations of input fea￾tures: process parameters (PP), kiln feed (KF), hot meal (HM), and clinker oxides (CO). Values in parentheses represent MAPE (%) for alite (green), ferrite (blue), and belite (red) predictions. b-g, Per… view at source ↗
Figures from the paper (1 more)
Figure 4
Figure 4. Figure 4: Feature attribution analysis of clinker phase predictions using SHAP. a-c, Hierarchical ranking of clinker oxide contributions to phase predictions for alite, belite, and ferrite, respectively. Bar lengths indicate mean absolute SHAP values (wt.%), representing average…

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

58 extracted references · 46 canonical work pages

  1. [1]

    Y . Kim, A. Hanif, M. Usman, M. J. Munir, S. M. S. Kazmi, S. Kim, Slag waste incorporation in high early strength concrete as cement replacement: Environmental impact and influence on hydration & durability attributes, Journal of Cleaner Production 172 (2018) 3056–3065, publisher: Elsevier. URL https://www.sciencedirect.com/science/article/pii/S0959652617327786

    work page 2018

  2. [2]

    D. K. Panesar, D. Kanraj, Y . Abualrous, Effect of transportation of fly ash: Life cycle assessment and life cycle cost analysis of concrete, Cement and Concrete Composites 99 (2019) 214–224, publisher: Elsevier. URL https://www.sciencedirect.com/science/article/pii/S0958946518308643

    work page 2019

  3. [3]

    Cancio D´ıaz, S

    Y . Cancio D´ıaz, S. S ´anchez Berriel, U. Heierli, A. R. Favier, I. R. S ´anchez Machado, K. L. Scrivener, J. F. Mar- tirena Hern´andez, G. Habert, Limestone calcined clay cement as a low-carbon solution to meet expanding cement demand in emerging economies, Development Engineering 2 (2017) 82–91. doi:10.1016/j.deveng.2017. 06.001. URL https://www.scienc...

    work page doi:10.1016/j.deveng.2017 2017

  4. [4]

    M. Zaki, S. Sharma, S. K. Gurjar, R. Goyal, Jayadeva, N. M. A. Krishnan, Cementron: Machine learning the alite and belite phases in cement clinker from optical images, Construction and Building Materials 397 (2023) 132425. doi:10.1016/j.conbuildmat.2023.132425. URL https://www.sciencedirect.com/science/article/pii/S0950061823021414

    arXiv 2023

  5. [5]

    Negash, G

    M. Negash, G. Alemu, Neural network based data-driven predictor: Case study on clinker quality prediction, Zede Journal 29 (2012) 27–36. URL https://www.ajol.info/index.php/zj/article/view/92044

    work page 2012

  6. [6]

    Sharma, P

    P. Sharma, P. N. Sheth, S. Sen, Aspen plus simulation of an inline calciner for white cement production with a fuel mix of petcoke and producer gas, Energy 282 (2023) 128892, publisher: Elsevier. URL https://www.sciencedirect.com/science/article/pii/S0360544223022867

    work page 2023

  7. [7]

    Zhang, S.-X

    Y . Zhang, S.-X. Cao, S. Shao, Y . Chen, S.-L. Liu, S.-S. Zhang, Aspen Plus-based simulation of a cement calciner and optimization analysis of air pollutants emission, Clean Technologies and Environmental Policy 13 (3) (2011) 459–468. doi:10.1007/s10098-010-0328-y . URL http://link.springer.com/10.1007/s10098-010-0328-y

    work page doi:10.1007/s10098-010-0328-y 2011

  8. [8]

    Redjeba, K

    Y . Redjeba, K. Kaabeche-Djerafia, O. Saoulia, AspenPlus based simulation for waste heat recovery in cement industries, J New Technol Mater 8 (2019) 82–6. URL https://www.researchgate.net/profile/Youcef-Redjeb/publication/332100849_ AspenPlus_based_simulation_for_waste_heat_recovery_in_cement_industries/links/ 5c9fe26992851cf0aea2ba00/AspenPlus-based-simu...

    work page arXiv 2019

Show all 58 references
  1. [9]

    Rahman, M

    A. Rahman, M. G. Rasul, M. M. K. Khan, S. Sharma, Aspen Plus based simulation for energy recovery from waste to utilize in cement plant preheater tower, Energy Procedia 61 (2014) 922–927, publisher: Elsevier. URL https://www.sciencedirect.com/science/article/pii/S1876610214028264

    2014

  2. [10]

    K ¨a¨antee, R

    U. K ¨a¨antee, R. Zevenhoven, R. Backman, M. Hupa, Cement manufacturing using alternative fuels and the advan- tages of process modelling, Fuel processing technology 85 (4) (2004) 293–301, publisher: Elsevier. URL https://www.sciencedirect.com/science/article/pii/S0378382003002030

    2004

  3. [11]

    Arachchige, D

    U. Arachchige, D. Kawan, L.-A. Tokheim, M. C. Melaaen, Model Development for CO2 capture in the cement industry, International Journal of Modeling and Optimization 3 (6) (2013). URL https://www.researchgate.net/profile/Prem-Baboo/post/Can-anyone-help-me-to-find-a-proper-simula...

    2013

  4. [12]

    Meyer, A

    V . Meyer, A. Pisch, K. Penttil ¨a, P. Koukkari, Computation of steady state thermochemistry in rotary kilns: Ap- plication to the cement clinker manufacturing process, Chemical Engineering Research and Design 115 (2016) 335–347

    2016

  5. [13]

    Mastorakos, A

    E. Mastorakos, A. Massias, C. D. Tsakiroglou, D. A. Goussis, V . N. Burganos, A. C. Payatakes, CFD predictions for cement kilns including flame modelling, heat transfer and clinker chemistry, Applied Mathematical Modelling 23 (1) (1999) 55–76, publisher: Elsevier. URL https://...

    1999

  6. [14]

    Z. Wang, T. R. Wang, M. Z. Yuan, H. Wang, Dynamic model for simulation and control of cement rotary kilns, System Simulation 20 (19) (2008) 5131–5135

    2008

  7. [15]

    Darabi, A mathematical model for cement kilns, PhD Thesis, University of British Columbia (2007)

    P. Darabi, A mathematical model for cement kilns, PhD Thesis, University of British Columbia (2007). URL https://open.library.ubc.ca/soa/cIRcle/collections/ubctheses/831/items/1.0080783

    2007

  8. [16]

    Sadighi, M

    S. Sadighi, M. Shirvani, A. Ahmad, Rotary cement kiln coating estimator: Integrated modelling of kiln with shell temperature measurement, The Canadian Journal of Chemical Engineering 89 (1) (2011) 116–125. doi: 15 10.1002/cjce.20365. URL https://onlinelibrary.wiley.com/doi/10....

    2011 doi

  9. [17]

    N.-O. E. Moses, S. B. Alabi, Predictive model for cement clinker quality parameters, Journal of Materials Science and Chemical Engineering 4 (7) (2016) 84–100, publisher: Scientific Research Publishing. URL https://www.scirp.org/journal/paperinformation.aspx?paperid=69058

    2016

  10. [18]

    A. Dean, D. V oss, D. Dragulji´c, Response Surface Methodology, in: Design and Analysis of Experiments, Springer International Publishing, Cham, 2017, pp. 565–614, series Title: Springer Texts in Statistics. doi:10.1007/ 978-3-319-52250-0_16 . URL http://link.springer.com/10.1...

    2017 doi

  11. [19]

    N. A. Krishnan, R. Bhattoo, H. Kodamana, Machine Learning for Materials Discovery: Numerical Recipes and Practical Applications, Springer Nature, 2024

    2024

  12. [20]

    Z. Li, J. Yoon, R. Zhang, F. Rajabipour, W. V . Srubar III, I. Dabo, A. Radli ´nska, Machine learning in concrete science: applications, challenges, and best practices, npj Computational Materials 8 (1) (2022) 127, publisher: Nature Publishing Group UK London. URL https://www....

    2022

  13. [21]

    A. M. Ali, J. D. Tabares, M. W. McGinley, A machine learning approach for clinker quality prediction and nonlinear model predictive control design for a rotary cement kiln, Journal of Advanced Manufacturing and Processing 4 (4) (2022) e10137. doi:10.1002/amp2.10137. URL https:...

    2022 doi

  14. [22]

    K. M. Jablonka, C. Charalambous, E. Sanchez Fernandez, G. Wiechers, J. Monteiro, P. Moser, B. Smit, S. Garcia, Machine learning for industrial processes: Forecasting amine emissions from a carbon capture plant, Science Advances 9 (1) (2023) eadc9576. doi:10.1126/sciadv.adc9576...

    2023 doi

  15. [23]

    G. A. Lyngdoh, M. Zaki, N. A. Krishnan, S. Das, Prediction of concrete strengths enabled by missing data imputa- tion and interpretable machine learning, Cement and Concrete Composites 128 (2022) 104414

    2022

  16. [24]

    Lapeyre, T

    J. Lapeyre, T. Han, B. Wiles, H. Ma, J. Huang, G. Sant, A. Kumar, Machine learning enables prompt prediction of hydration kinetics of multicomponent cementitious systems, scientific reports 11 (1) (2021) 3922

    2021

  17. [25]

    Miyan, N

    N. Miyan, N. A. Krishnan, S. Das, Integrating data imputation and augmentation with interpretable machine learn- ing for efficient strength prediction of fly ash-based alkali-activated concretes, Journal of Building Engineering 98 (2024) 111248

    2024

  18. [26]

    J. J. Monta ˜no Moreno, A. Palmer Pol, A. Ses ´e Abad, Using the R-MAPE index as a resistant measure of forecast accuracy, Psicothema (25.4) (2013) 500–506. doi:10.7334/psicothema2013.23. URL https://doi.org/10.7334/psicothema2013.23

    2013 doi

  19. [27]

    Note on regression and inheritance

    Pearson: VII. Note on regression and inheritance... - Google Scholar. URL https://scholar.google.com/scholar_lookup?title=Notes%20on%20regression%20and% 20inheritance%20in%20the%20case%20of%20two%20parents&author=P.%20Karl&publication_ year=1895&pages=240-242

  20. [28]

    URL https://scikit-learn/stable/modules/generated/sklearn.preprocessing

    sklearn.preprocessing.StandardScaler. URL https://scikit-learn/stable/modules/generated/sklearn.preprocessing. StandardScaler.html

  21. [29]

    Bates, T

    S. Bates, T. Hastie, R. Tibshirani, Cross-Validation: What Does It Estimate and How Well Does It Do It?, Journal of the American Statistical Association (2023) 1–12doi:10.1080/01621459.2023.2197686. URL https://www.tandfonline.com/doi/full/10.1080/01621459.2023.2197686

    2023

  22. [30]

    URL https://scikit-learn/stable/modules/generated/sklearn.model_selection

    sklearn.model selection.GridSearchCV. URL https://scikit-learn/stable/modules/generated/sklearn.model_selection. GridSearchCV.html

  23. [31]

    S. M. Lundberg, S.-I. Lee, A unified approach to interpreting model predictions, Advances in neural information processing systems 30 (2017). URL https://proceedings.neurips.cc/paper/2017/hash/8a20a8621978632d76c43dfd28b67767-Abstract. html

    2017

  24. [32]

    L. S. Shapley, A value for n-person gamesPublisher: Princeton University Press Princeton (1953). URL https://books.google.com/books?hl=en&lr=&id=Pd3TCwAAQBAJ&oi=fnd&pg= PA307&dq=+Shapley,L.S.AValueforn-PersonGames.(1952).&ots=gunTLcejtX&sig= rg59jZ2kwcR7N0K5QL140Ri3Z_s

    1953

  25. [33]

    K. A. Waris, S. J. Fayaz, A. H. Reddy, B. M. Basha, Pseudo-static slope stability analysis using explainable machine learning techniques, Natural Hazards (2024) 1–33

    2024

  26. [34]

    Bhattoo, S

    R. Bhattoo, S. Bishnoi, M. Zaki, N. M. A. Krishnan, Understanding the compositional control on electrical, me- chanical, optical, and physical properties of inorganic glasses with interpretable machine learning, Acta Materialia 242 (2023) 118439. doi:10.1016/j.actamat.2022.118...

    2023

  27. [35]

    Karimipour, S

    A. Karimipour, S. A. Bagherzadeh, A. Taghipour, A. Abdollahi, M. R. Safaei, A novel nonlinear regression model of SVR as a substitute for ANN to predict conductivity of MWCNT-CuO/water hybrid nanofluid based on empirical data, Physica A: Statistical Mechanics and its Applicati...

    2019 doi

  28. [36]

    Martinez-Ramon, J

    M. Martinez-Ramon, J. L. Rojo-Alvarez, G. Camps-Valls, J. Munoz-Mari, n. Navia-Vazquez, E. Soria-Olivas, A. R. Figueiras-Vidal, Support Vector Machines for Nonlinear Kernel ARMA System Identification, IEEE Transactions on Neural Networks 17 (6) (2006) 1617–1622, conference Nam...

    2006

  29. [37]

    W. Ni, S. K. Tan, W. J. Ng, S. D. Brown, Moving-Window GPR for Nonlinear Dynamic System Modeling with Dual Updating and Dual Preprocessing, Industrial & Engineering Chemistry Research 51 (18) (2012) 6416–6428, publisher: American Chemical Society. doi:10.1021/ie201898a. URL ht...

    2012 doi

  30. [38]

    W. Ni, S. K. Tan, W. J. Ng, Recursive GPR for nonlinear dynamic process modeling, Chemical Engineering Journal 173 (2) (2011) 636–643. doi:10.1016/j.cej.2011.08.021. URL https://www.sciencedirect.com/science/article/pii/S1385894711009491

    2011 doi

  31. [39]

    Tealab, H

    A. Tealab, H. Hefny, A. Badr, Forecasting of nonlinear time series using ANN, Future Computing and Informatics Journal 2 (1) (2017) 39–47. doi:10.1016/j.fcij.2017.05.001. URL https://www.sciencedirect.com/science/article/pii/S2314728817300144

    2017 doi

  32. [40]

    James, D

    G. James, D. Witten, T. Hastie, R. Tibshirani, An Introduction to Statistical Learning: with Applications in R, Springer Texts in Statistics, Springer US, New York, NY , 2021.doi:10.1007/978-1-0716-1418-1 . URL https://link.springer.com/10.1007/978-1-0716-1418-1

    2021 doi

  33. [41]

    C. M. Bishop, N. M. Nasrabadi, Pattern recognition and machine learning, V ol. 4, Springer, 2006. URL https://link.springer.com/book/9780387310732

    2006

  34. [42]

    Ghojogh, M

    B. Ghojogh, M. Crowley, The Theory Behind Overfitting, Cross Validation, Regularization, Bagging, and Boosting: Tutorial, arXiv:1905.12787 [cs, stat] (May 2023). URL http://arxiv.org/abs/1905.12787

    1905 arXiv

  35. [43]

    Anoop Krishnan, S

    N. Anoop Krishnan, S. Mangalathu, M. M. Smedskjaer, A. Tandia, H. Burton, M. Bauchy, Predicting the disso- lution kinetics of silicate glasses using machine learning, Journal of Non-Crystalline Solids 487 (2018) 37–45. doi:10.1016/j.jnoncrysol.2018.02.023. URL https://linkingh...

    2018 doi

  36. [44]

    H. Zou, T. Hastie, Regularization and Variable Selection Via the Elastic Net, Journal of the Royal Statistical Society Series B: Statistical Methodology 67 (2) (2005) 301–320. doi:10.1111/j.1467-9868.2005.00503.x. URL https://doi.org/10.1111/j.1467-9868.2005.00503.x

    2005

  37. [45]

    Breiman, Bagging predictors, Machine Learning 24 (2) (1996) 123–140

    L. Breiman, Bagging predictors, Machine Learning 24 (2) (1996) 123–140. doi:10.1007/BF00058655. URL http://link.springer.com/10.1007/BF00058655

    1996 doi

  38. [46]

    Bartlett, Y

    P. Bartlett, Y . Freund, W. S. Lee, R. E. Schapire, Boosting the margin: A new explanation for the e ffectiveness of voting methods, The annals of statistics 26 (5) (1998) 1651–1686, publisher: Institute of Mathematical Statistics. URL https://projecteuclid.org/journals/annals...

    1998

  39. [47]

    A. Liaw, M. Wiener, Classification and Regression by randomForest 2 (2002)

    2002

  40. [48]

    T. Chen, C. Guestrin, Xgboost: A scalable tree boosting system, in: Proceedings of the 22nd acm sigkdd interna- tional conference on knowledge discovery and data mining, 2016, pp. 785–794

    2016

  41. [49]

    Guardiani, E

    C. Guardiani, E. Soranzo, W. Wu, Time-dependent reliability analysis of unsaturated slopes under rapid drawdown with intelligent surrogate models, Acta Geotechnica (2022) 1–26

    2022

  42. [50]

    C. E. Rasmussen, C. K. Williams, et al., Gaussian processes for machine learning, V ol. 1, Springer, 2006

    2006

  43. [51]

    T. I. Barry, F. P. Glasser, Calculations of Portland cement clinkering reactions, Advances in Cement Research 12 (1) (2000) 19–28, publisher: ICE Publishing. doi:10.1680/adcr.2000.12.1.19. URL https://www.icevirtuallibrary.com/doi/full/10.1680/adcr.2000.12.1.19 17 Industrial-s...

    2000 doi

  44. [52]

    Ntree bootstraped samples are drawn from the training data

  45. [53]

    Unlike a regular bagging scenario, where Mrandom = n, the RF algorithm randomly samples Mrandom features, usually 2√n or log2(n), which are used to split each node

    For each bootstraped sample, an unpruned decision tree is grown. Unlike a regular bagging scenario, where Mrandom = n, the RF algorithm randomly samples Mrandom features, usually 2√n or log2(n), which are used to split each node. The best split among the sampled features is de...

  46. [54]

    At each bootstrap iteration, the data points not in the bootstraped sample, called the out-of-bag (OOB) data, are predicted by the tree grown with that bootstraped sample

  47. [55]

    yt f∗ # ∼ N 0,

    OOB predictions are aggregated to estimate the OOB error rate. Appendix B.5. XGBoost XGBoost is a boosting-based ensembling method wherein base learners (decision trees) are trained in succession to minimize the errors of the preceding trees. The training continues until a spe...

  48. [56]

    k(x∗ 1, x∗ n) (x∗ 2, x∗

  49. [57]

    k(x∗ 2, x∗ n) ... ... ... ... (x∗ n, x∗

  50. [58]

    k(x∗ n, x∗ n)  (B.24) and likewise forK(Xt, X∗) and K(X∗, X∗). Using conditional distribution corresponding to EqnB.23, the predictive equations for GPR can be written as [50]: f∗|Xt, yt, X∗∼ N  m( f∗)|{z} Predective mean , k( f∗)|{z} Predective cova...

    1930

Pith tools

Reviewed August 11, 2026 · model on record in the stance chip above.