1. Modelling and validation of liquefaction potential index of fine-grained soils using ensemble learning paradigms

Contributors : Sufyan Ghani, Sanjog Chhetri Sapkota, Raushan Kumar Singh, Abidhan Bardhan, Panagiotis G. Asteris

Highlights : 

1. Employed ensemble machine learning to predict the liquefaction potential index of fine-grained soil. 

2. Higher prediction accuracy of 99%, which makes it feasible for early assessments of liquefaction susceptibility.

3. Verified using a  Chi-Chi earthquake validation set for the developed model. 

Status : 

Published in  Soil Dynamics and Earthquake Engineering ( Elsevier)

Link : 

https://doi.org/10.1016/j.soildyn.2023.108399 

Link to Pdf : 

Modelling and validation of liquefaction potential index of fine-grained soils using ensemble learning paradigms

3. Optimized Machine Learning Models for Prediction of Effective Stiffness of Rectangular Reinforced Concrete Column Sections 

Contributors:  Sanjog Chhetri Sapkota , Sourav Das , Prasenjit Saha 

Highlights : 

1. The primary objective of the present study is to predict the effective stiffness of rectangular RC columns using hybrid ensemble ML models combined with TPE-based Bayesian optimization. 

2. It was prepared by computing the effective stiffness ratio (𝑟 = 𝐸𝐼𝑒⁄𝐸𝐼𝑔) of columns from a set of RC frame buildings designed as per unified performance-based design (UPBD) methodology.Inputs include the amount of longitudinal rebar (𝑝𝑡), the axial load (𝑀), and the column section dimensions (𝐷𝑥,𝐷𝑦). Outputs are the effective stiffness ratios in both orthogonal directions (𝑟𝑥, 𝑟𝑦).

3. The results of the present study exhibit that the CatBoost model outperforms other considered machine learning models with 𝑅2 value of 0.9921 and 0.9966. 

4. A sensitivity analysis based on SHAP has also been implemented to establish the correlation between input parameters and output parameters.  

Status : 

Published  in Structures, Elsevier, I.F 4.1 

Link : https://doi.org/10.1016/j.istruc.2024.106155  

4. Leveraging Explainable Hybrid Machine learning for the Prediction of Interfacial Bond Strength between Normal Concrete and Ultra-High-Performance Concrete 

Contributors: Sanjog Chhetri Sapkota, Sagar Sapkota, Tushar Bansal

Highlights : 

1.       This paper highlights the use of hybrid algorithms (ensemble ML algorithms and TPE-based Bayesian optimization) with better predictability.

2.    Incorporation of a hybrid algorithm with Nested Cross validation for robust prediction in unseen data.       

3.    Highlights the factors affecting the bond strength of UHPC-NSC, explaining local and global interperability using SHAP analysis.

4.    Addition of the reverse design method using Shap results for better insights into split tensile and slant shear data for UHPC-NSC bond strength. 

Status : 

Published in Arabian Journal for Science and Technology , Springer

Link: https://doi.org/10.1007/s13369-024-09690-3   

5. A Generalised Explainable Approach to Predict the Hardened Properties of Self-Compacting Geopolymer Concrete Using Machine Learning Techniques

Contributors: :Endow Ayar Mazumder, Sanjog Chhetri Sapkota, Sourav Das, Prasenjit Saha, and Pijush Samui

Highlights : 

1. Employed hybrid machine learning to predict the hardened properties of self-compacting geopolymer concrete for the laboratory experiments . 

2. Higher prediction accuracy of more than  99% for compressive strength, flexural strength, and split tensile strength.

3. Used  shap explainable method for the model in giving the better decision with optimized results. 

Status : 

Published, Computers and Concrete (Techno press (SCI-indexed), I.F 4.1)

Link : https://doi.org/10.12989/cac.2024.34.3.279 

6. Analyzing the behavior of geopolymer concrete (GPC) with different novel machine-learning algorithms 

Contributors: : Sanjog Chhetri Sapkota, Dipak Dahal, Ajay Yadav, Dipak Dhakal, Satish Paudel 

Highlights : 

1. Studied the application of ensemble machine learning (ML) models to predict the fresh and hardened properties of self-compacting concrete (SCC).

2. Considered six input variables: cement dosage, fly ash, water–powder ratio, coarse aggregate, fine aggregate, and super-plasticizer, with output variables: slump flow, L-box ratio, V-funnel, and compressive strength (Fc28).

3. Models used: Gradient Boosting (GBR), Adaboost, LightGBM, XGBoost, and CatBoost.

4. XGBoost achieved the highest accuracy for predicting slump flow (R² = 0.9941) and V-funnel (R² = 0.9981).

5. CatBoost performed best for L-box ratio (R² = 0.9926) and compressive strength (R² = 0.9941).

6. Sensitivity analysis revealed the influence of input ingredients on fresh and hardened SCC properties, providing insights for mix optimization.

7. Highlights the effectiveness of XGBoost and CatBoost as reliable tools for predicting SCC properties accurately.

Status : 

Accepted (Yet to publish in Practice Periodical on Structural Design and Construction , ASCE (SCI-indexed) 

Link :

7.Prediction of Strength and Failure Modes in RC Beam-Column Joints using  Ensemble Machine Learning Techniques 

Contributors: : Prasenjit Saha, Sanjog Chhetri Sapkota, Sourav Das 

Highlights : 

1. Focused on beam-column joints, critical structural elements that influence structural behavior during seismic excitation.

2. Evaluated response properties, including failure modes and ultimate shear capacity of exterior beam-column joints.

3. Developed prediction models using ensembled machine learning approaches: Decision Tree, Random Forest, AdaBoost, CatBoost, and LightGBM.

4. The CatBoost model demonstrated superior performance with high correlation coefficients for predicting load-carrying capacity (0.9836) and shear strength (0.9978).

5. The CatBoost algorithm also achieved the highest accuracy in classifying the failure modes of beam-column joints.

6. Highlighted the potential of CatBoost as an efficient tool for predicting shear strength and failure modes, reducing the time required for conventional methods.

Status : Accepted (Published, Structural Engineering International , Taylor and Francis  (SCI-indexed)

Link :    https://doi.org/10.1080/10168664.2024.2403593 

 8. Explainable hybridized ensemble machine learning for the prognosis of the compressive strength of recycled plastic-based sustainable concrete with experimental validation

Contributors: Sanjog Chhetri Sapkota , Ajay Yadav,Ajaya Khatri, Tushar Singh and Dipak Dahal 

Highlights : 

1. Investigated the use of recycled aggregate concrete to reduce environmental impact while addressing its effect on compressive strength (CS).

2. Applied Random Forest and XGB models with optimization algorithms Krill Herd (KH) and Leopard Seal Algorithm (LSA).

3. XGB-LSA achieved R² = 0.97 (training) and R² = 0.95 (testing), demonstrating exceptional performance.

4. Tenfold cross-validation ensured accurate predictions and minimized overfitting risks.

5. SHAP analysis identified critical components: binder (550 kg/m³), recycled plastic aggregate (<25 kg/m³), and W/B ratio (0.3–0.35).

6. Validation datasets exhibited 94% accuracy, confirming model reliability.

7. Developed a GUI for real-time application using the best-performing model.

8. Demonstrated the potential of machine learning for sustainable construction and informed decision-making.

Status : 

Published  in Multiscale and Multidisciplinary Modeling, Experiments and Design, Springer

Link : https://doi.org/10.1007/s41939-024-00567-4  

9. Influence of Fiber Content on Autogenous Shrinkage and Mechanical Properties of High-Strength Engineered Cementitious Composites 

Contributors:  Ajad Shrestha, Sanjog Chhetri Sapkota, Bijay Bhatt, Md Ahatasamul Hoque 

Highlights : 

1. Reduced autogenous shrinkage in HS-ECC using waste iron sand and varying PE fiber content (1.2%, 1.5%, 1.8%).

2. Achieved a maximum shrinkage reduction of 40.76% at 1.8% fiber content.

3. Enhanced compressive strength up to 118.4 MPa and tensile strength to 13.43 MPa with tensile strain of 7.59%.

4. Elastic modulus remained stable across all fiber levels, exceeding 40 GPa.

5. Promoted the use of waste iron sand and fiber for sustainable HS-ECC production.

6. Suitable for civil engineering applications with improved mechanical properties and reduced shrinkage.

Status : 

Ongoing Review  in Journal of Sustainable Cement-Based Materials, Taylor and Francis 

Link : 

10. A Paradigm Shift in Corrosion Inhibition using Botanical Extracts: From Conventional Methods to Advanced Methods for Reinforcing Steel  

Contributors: Sanjog Chhetri Sapkota, Dipak Dahal, Ajay Yadav, Dipak Dhakal, Ram Kumar Sharma, Gaurav Saini

Highlights : 

1. Explores the evolution of Corrosion Inhibitors (CIs) in RCC, transitioning from conventional chemical methods to sustainable and advanced alternatives.

2. Focuses on Green Corrosion Inhibitors (GCIs), specifically Botanical Corrosion Inhibitors (BCIs), and their inhibition mechanisms.

3. Reviews various BCI tests, including preparation and extraction methods, phytochemical identification, inhibitor efficiency, surface morphology, and computational modeling.

4. Discusses advanced simulation and data-driven techniques like DFT, Molecular Dynamics Simulations (MDS), Monte Carlo Simulation (MCS), QSAR, Machine Learning (ML), and Deep Learning (DL) for corrosion analysis.

5. Provides a comparative summary of GCIs, detailing phytochemical content, corrosive medium, concentration, and the relationship between inhibition efficiency, temperature, concentration, and immersion time.

6. Highlights past, present, and future approaches for corrosion prevention in RCC, emphasizing sustainable practices. 

Status : 

Ongoing Review in Innovative Infrastructures Solutions

Link : SSRN link : https://dx.doi.org/10.2139/ssrn.4963462 

11. Techno-economic analysis of hydrogen production from waste plastics and storage plant in the context of Japan

Contributors:  Bishwash Paneru , Biplov Paneru , Sanjog Chhetri Sapkota ,  Dhiraj Kumar Mandal ,  Prem Giri 

Highlights : 

1. Converts non-biodegradable plastics into hydrogen to address Japan's plastic waste issue.

2. Uses DWSIM for plastic-to-hydrogen conversion and Aspen Plus for hydrogen compression.

3. It produces 1,449,792 kg/year of hydrogen from plastics such as PET, PVC, PE, PP, and PS.

4. Operates at 600 °C and 3 bar, processing 100 kg/h of plastic to yield 7.098 kg/h of hydrogen.

5. Economic analysis: CAPEX of 143.5 million JPY and OPEX of 29.7 million JPY/year.

6. Levelized Cost of Hydrogen (LCOH): 8.874–19.82 USD/kg.

7. Conducts sensitivity and uncertainty analyses for LCOH variability.

8. Combines waste management with clean energy production for sustainable development.

Status : 

Published  in International Journal of Hydrogen Energy, Elsevier, I.F 8.1 

Link : https://doi.org/10.1016/j.ijhydene.2024.11.134 

12. Advanced predictive modelling of electrical resistivity for geotechnical and geo-environmental applications using machine learning techniques 

Contributors:  Soumitra Kumar Kundu , Ashim Kanti Dey , Sanjog Chhetri Sapkota , Prasenjit Debnath , Prasenjit Saha , Arunava Ray , Manoj Khandelwal 

Highlights : 

1. Utilizes Electrical Resistivity (ER) for non-invasive, economical, and rapid subsurface investigations.

2. Conducted 2772 ER tests on seven soil types with varying temperature, density, and water content.

3. Developed predictive models using SVR, ANN, and XGB, with XGB achieving the best performance (R² = 0.99).

4. Parametric study identified the influence of input parameters (temperature, density, moisture) on ER.

5. Error analysis validated the robustness and reliability of the XGB model as a substitute method for ER prediction. 

Status : 

Published  in Journal of GeoPhysics, Elsevier 

Link : https://doi.org/10.1016/j.jappgeo.2024.105557 

13. Experimental Study with Machine Learning Validation for Analyzing the Influence of  Oxide Ratio of Alumina Silicate Precursor for Development of Geopolymer 

Contributors:  Ashwin N Raut, Anant Lal Murmu, Sanjog Chhetri Sapkota, Sourav Das, Prasenjit Saha  

Highlights : 

1. Analyzed the impact of oxide ratios (Si/Al, Al/Na, Si/Na, Na/H2O) on the physical and mechanical properties of geopolymer concrete.

2. Prepared six FA-based and two GGBS-based geopolymer mixes with NaOH molarities of 10, 12, and 14.

3. Utilized machine learning models (Random Forest, Gradient Boosting, AdaBoost, and Stacking) to predict compressive strength.

4. Stacking model outperformed others with high R² values: 0.977 (3-day), 0.956 (7-day), and 0.951 (28-day).

5. SHAP analysis identified Na/Si ratio and Setting Time (ST) as the most influential factors for early strength prediction.

6. Provides a reliable predictive method for compressive strength, facilitating the integration of geopolymers into the construction industry.

Status : 

(Accepted in (Construction and Building Materials, Elsevier )

Link : 

14. Prediction of bearing capacity of geogrid-reinforced sand over stone columns in soft clay

Contributors: Amit Saha, Sanjog Chhetri Sapkota, Prasenjit Saha, Prasenjit Debnath, Suman Hazari, Sourav Das, Pijush Samu 

Highlights : 

1. Developed a machine learning model to predict the ultimate bearing capacity (qrs) of geogrid-reinforced sandy beds on vertical stone columns in soft clay.

2. Utilized a dataset of 245 experimental observations for model training and evaluation.

3. Ensemble techniques like KNN, Random Forest, and XGBoost were employed, with XGBoost achieving the best performance (R² = 0.9947).

4. Key influencing parameters include geogrid properties, column spacing, column diameter, and soil strength.

5. SHAP analysis revealed the significance of each parameter in predicting bearing capacity, enhancing model explainability.

6. Demonstrates the potential of XGBoost as a reliable and cost-effective tool for practical geotechnical applications.

Status : 

Published  in Proceedings of the Institution of Civil Engineers-Ground Improvement , Emerald Publishing Limited 

Link : https://doi.org/10.1680/jgrim.24.00032 

15. Hybrid machine learning model to predict the mechanical properties of ultra-high-performance concrete (UHPC) with experimental validation

Contributors:  Ajad Shrestha, Sanjog Chhetri Sapkota

Highlights : 

1. Investigates the use of Ultra-High-Performance Concrete (UHPC) incorporating waste cementitious materials for enhanced mechanical strength, durability, and environmental sustainability.

2. Developed predictive models using Random Forest (RF) and XGBoost (XGB), optimized with Pelican Optimization Algorithm (POA) and Walrus Optimization Algorithm (WOA) for hyperparameter tuning.

3. XGB-POA achieved the highest accuracy with an R² > 0.96 in the testing set, demonstrating its reliability.

4. Employed ten-fold cross-validation to ensure model robustness and mitigate overfitting.

5. SHAP analysis identified age, fiber content, cement, and silica fume (SF) dosage as the most influential features for UHPC performance.

6. Validated the models using experimental datasets, achieving over 95% accuracy.

7. Provides actionable insights into feature relationships, supporting sustainable practices in UHPC applications.

Status : 

Published  in Asian Journal of Civil Engineering, Springer

Link : https://doi.org/10.1007/s42107-024-01109-6   

16. Exploring the potential of Himalayan Giant Nettle fiber and supplementary cementitious materials for sustainable concrete development

Contributors:  Ajaya Subedi, Bhum Bahadur Thapa, Ashish Poudel, Binaya Adhikari, Binod Khadka, Samrat Poudel & Sanjog Chhetri Sapkota

Highlights : 

1. Evaluated Himalayan Giant Nettle (HGN) fiber as a novel natural reinforcement in M30 concrete, alongside Fly Ash (FA) and Rice Husk Ash (RHA) as supplementary cementitious materials (SCMs).

2. FA and RHA replacements at 20% by weight enhanced compressive strength (CS), flexural strength (FS), and split tensile strength (STS).

3. 1% HGN fiber content optimally increased CS, FS, and STS by 16.2%, 33.33%, and 36.90%, respectively.

4. Exceeding 1% HGN fiber reduced workability and strength, highlighting the importance of optimal content.

5. HGN fiber’s ability to bridge cracks, reduce stress concentration, and improve flexibility contributed to strength improvements.

6. Offers a low-cost, renewable, and lightweight solution, enhancing the sustainability of concrete production.

7. Confirms HGN fiber as a promising eco-friendly alternative for concrete reinforcement, advancing sustainable construction practices.

Status : 

Published  in Asian Journal of Civil Engineering, Springer

Link : https://doi.org/10.1007/s42107-024-01211-9  

17.  Comparative nonlinear analysis of confined masonry walls: simplified micro-modeling versus finite element modelling in ETABS

Contributors: G.Panth, A.Shrestha, Sanjog Chhetri Sapkota 

Highlights : 

1. This study highlights the development of a simplified micro-modeling technique for the nonlinear analysis of confined masonry (CM) walls, offering a new approach to structural analysis.

2. Incorporation of this micro-modeling technique alongside conventional finite element modeling (FEM) in ETABS for a robust comparison using identical modeling parameters.

3. Highlights the application of the concrete damage plasticity (CDP) model and truss elements to represent brick and concrete tie elements, providing insights into structural behavior at the macro level.

4. A full-scale CM building model was added for seismic performance assessment, validating the efficacy of both modelling approaches in earthquake-prone conditions.

5. Comparative analysis of numerical results from both modeling techniques with experimental results from existing literature, establishing the reliability of the simplified micro-modeling technique for future use in seismic resilience studies of CM structures.

Status : (Published, Asian Journal of Civil Engineering, Springer)

Link : https://doi.org/10.1007/s42107-024-01123-8  

18. Prediction of Seismic Response and Fundamental Period of Vibration of Low-Rise Reinforced Moment Resisting Frame Building with Masonry Infill Using Hybridized Machine Learning 

Contributors:  Ajaya Khatri, Ashish Thapa, Sanjog Chhetri Sapkota 

Highlights : 

1. Proposed a hybridized machine learning (ML) approach to predict seismic response and fundamental period (T) of RC-MRFs, considering masonry infill walls.

2. Modeled and analyzed 600 3D low-rise RC-MRFs in OpenSees with ten input parameters and four outputs (DI, DR, SH, T).

3. Developed prediction models using XGB and RF, optimized with Brown-Bear Optimization Algorithm (BOA) and Whale Optimization Algorithm (WOA).

4. Achieved high R² values (0.976–0.997 for training, 0.928–0.988 for testing), outperforming empirical equations for T prediction.

5. XGB-BOA excelled in predicting DI, SH, and T, while RF-BOA performed better for DR.

6. Presented results via an interactive web-based GUI for real-time application, aiding engineers and researchers.

Status : 

On going Review  in (Soil Dynamics and Earthquake Engineering, Elsevier)

Link : https://dx.doi.org/10.2139/ssrn.5000857 

19. Prediction of autogenous shrinkage in ultra-high-performance concrete (UHPC) using hybridized machine learning

Contributors:  Md Ahatasamul Hoque, Ajad Shrestha, Sanjog Chhetri Sapkota, Asif Ahmed & Satish Paudel

Highlights : 

1. Explored hybridized machine learning (ML) techniques to predict autogenous shrinkage (AS) in ultra-high-performance concrete (UHPC).

2. Base algorithms include Random Forest (RF), Extra Tree Regressor (ETR), Light Gradient Boosting Machine (LGBM), and XGBoost.

3. Proposed a novel Sparrow Search Algorithm (SSA)-XGBoost hybrid, achieving the best performance with R² = 0.91 and RMSE = 79.2 on the testing set.

4. Five-fold cross-validation ensures the model's robustness and minimizes overfitting risks.

5. Comparison of RMSE for other models: XGB (102.22), LGBM (108.38), ETR (87.42), and RF (98.57).

6. Key influential features identified through model explainability: curing relative humidity (CRH), steel fiber content (SFS), and sand.

7. Provides a comprehensive tool to understand factors influencing AS, aiding in better decision-making for UHPC design.

Status : 

Published  in Asian Journal of Civil Engineering, Springer

Link : https://doi.org/10.1007/s42107-024-01212-8  

20. Prediction of fresh and hardened properties of self-compacting concrete using ensemble soft learning techniques 

Contributors: Prasenjit Saha, Sanjog Chhetri Sapkota , Sourav Das, Naveen Kwatra 

Highlights : 

1. Studied the application of ensemble machine learning (ML) models to predict the fresh and hardened properties of self-compacting concrete (SCC).

2. Considered six input variables: cement dosage, fly ash, water–powder ratio, coarse aggregate, fine aggregate, and super-plasticizer, with output variables: slump flow, L-box ratio, V-funnel, and compressive strength (Fc28).

3. Models used: Gradient Boosting (GBR), Adaboost, LightGBM, XGBoost, and CatBoost.

4. XGBoost achieved the highest accuracy for predicting slump flow (R² = 0.9941) and V-funnel (R² = 0.9981).

5. CatBoost performed best for L-box ratio (R² = 0.9926) and compressive strength (R² = 0.9941).

6. Sensitivity analysis revealed the influence of input ingredients on fresh and hardened SCC properties, providing insights for mix optimization.

7. Highlights the effectiveness of XGBoost and CatBoost as reliable tools for predicting SCC properties accurately. 

Status : 

Published  in Multiscale and Multidisciplinary Modeling, Experiments and Design, Springer

Link : https://doi.org/10.1007/s41939-024-00423-5  

21. Prediction of split tensile strength of recycled aggregate concrete leveraging explainable hybrid XGB with optimization algorithm

Contributors:  Sanjog Chhetri Sapkota , Sagar Sapkota , Gaurav Saini 

Highlights : 

1. Utilizes Recycled Aggregate Concrete (RAC) to reduce waste and environmental impact.

2. Predicts split tensile strength (STS) using a hybridized XGB model.

3. Best performance achieved with XGB-PSO (R²: 0.999 training, 0.960 testing).

4. Key factors: Cement (300–500 kg/m³), M-RCA size (10–20 mm), Water (180–200 kg/m³).

5. SHAP analysis provides model explainability and validates experimental findings.

6. 10-fold cross-validation ensures robust and reliable predictions.

Status : 

Published  in Multiscale and Multidisciplinary Modeling, Experiments and Design, Springer

Link : https://doi.org/10.1007/s41939-024-00480-w 

22. Discharge Determination of Rectangular  Sharp Crested Weirs Using Machine Learning Models

Contributors: Mrimoy Dhar , Sanjog Chhetri Sapkota ,  Prasenjit Saha

Highlights : 

1.    This paper highlights the use of hybrid (ensemble ML algorithms and the comparison of standalone and ensemble algorithm. 

2.    Highlights the factors affecting the coefficient  of discharge with SHAP based Feature importance.

3. Comparison of different existing empirical models and the better performing machine learning models. 

4. The additional insights on the variation of discharge coefficient with weir contraction ratio  and Variation of discharge coefficient with the ratio h/P using machine learning results. 

Status : 

Reviewing in  (Flow Measurement and instrumentation, Elsevier, I.F 2.2)

24. Influence of cross‑girder spacing on RC T‑beam bridges: a comparative theoretical and finite element approach 

Contributors:  Ajaya Subedi,  Binaya Adhikari, Ashish Poudel, Binod Khadka, Samrat Poudel, Sanjog Chhetri Sapkota

Highlights : 

1. Cross girders are essential elements in T-beam bridges, providing lateral stiffening to longitudinal girders and reducing exterior girder torsion.

2. Studied the effects of varying cross-girder spacing on key structural parameters: Shear Force (SF), Maximum Displacement, Bending Moment (BM), and Torsion.

3. Developed multiple models in Finite Element Analysis (FEA) software with cross girders ranging from 0 to 10, keeping the span and width constant.

4. Loadings were applied as per IRC code, with dead load and live load combinations analyzed.

5. Odd numbers of cross girders were found more effective for live load distribution.

6. Observed that BM and deflection decrease when cross girders increase from 2 to 3 but rise with further additions. The optimum number of cross girders for reducing BM and deflection is 3.

7. Compared distribution factors from FEA software and Courbon’s distribution factor, noting variations.

Status : 

Published  in Discover Civil Engineering , Springer

Link : https://doi.org/10.1007/s44290-024-00126-3 

25. Estimation of Reference Crop Evapotranspiration: A Comparative Analysis

Contributors: Umesh Kumar Das ,Sanjog Chhetri Sapkota, Prasenjit Saha, Sameer Arora

Highlights : 

1. Used several machine learning algorithms including standalone, bagging and boosting based ensemble for the prediction of the reference crop evapotranspiration in Jaipur, Rajasthan, India, which falls under arid regions of India. 

2. Higher prediction accuracy of more than  95% on peak seasons where evapotranspiration is seen higher. 

3. It captures the pattern of 100 years and  the factors  affecting evapotranspiration like precipitation, vapour pressure and average temperature. 

Status : 

Reviewing in Hydrogeology journal , Springer (I.F 2.8)

26. Prediction of Sustainable Concrete based on Palm Fibre and activated carbon using Boosting-Based based ensemble machine learning techniques 

Contributors: Sanjog Chhetri Sapkota , Musa Adamu , Prasenjit Saha, Sourav Das

Highlights : 

1.  This paper highlights the  use of Date Palm Fibre and activated carbon for as a sustainable materials. 

2.  Use of hybrid (ensemble ML algorithm) and the comparison of standalone and ensemble algorithm. 

3.  Highlights the factors affecting fresh and hardened properties using boosting based ensemble model.

4.  Comparison of different existing machine learning models. 

5. Shap Analysis like  feature importance , dependence plots etc were used for model's explainability.

Status : 

Reviewing in  ( Congent Engineering, Taylor and Francis)