A Comparative Analysis of Machine Learning Models in Prediction of Mortar Compressive Strength

Gayathri, R.; Rani, Shola Usha; Čepová, Lenka; Raj, C. Justin; Kalita, Kanak

doi:10.3390/pr10071387

Cited by 28 publications

(9 citation statements)

References 45 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…To predict the SCP values for MOF/ethanol working pairs, six commonly used ML algorithms, namely, the random forest (RF), light gradient boosting machine (LGBM), , extreme gradient boosting (XGB), gradient boosting regression (GBR), hist gradient boosting regression (HGBR), and gradient boosted regression tree (GBRT) taken from the scikit-learn package, were trained in this work with all of the descriptors mentioned. We randomly partitioned the data from HTCS into training and testing sets, allocating 80% of the data for model training and reserving 20% for testing purposes.…”

Section: Methodsmentioning

confidence: 99%

Performance Evaluation of Metal–Organic Frameworks in Adsorption Heat Pumps via Multiscale Modeling

Liang,

Li,

et al. 2024

ACS Sustainable Chem. Eng.

View full text Add to dashboard Cite

The adsorption heat pump (AHP) driven by low-grade thermal energy is a promising technology to reduce building energy consumption for sustainable energy. Using metal−organic frameworks (MOFs) as adsorbents has attracted widespread attention in AHPs due to their large capacity of working fluids, a stepwise adsorption isotherm that tends to possess outstanding equilibrium performance (i.e., coefficient of performance, COP). Nevertheless, the dynamic performance of MOFs in AHPs lacks a quick evaluation and screening strategy, especially for specific cooling power (SCP) that is equally important with COP during operation. Herein, multiscale modeling combining the molecular simulation and the mathematical simulation of AHPs was proposed to obtain the SCP and COP for a vast number of MOF-based working pairs with high efficiency. Structure−property relationship obtained from the high-throughput computational screening of 1072 MOFs indicated that relatively low density (<1 kg/m 3 ), large pore size (>10 Å), and a relatively high void fraction (∼0.6) benefited the improvement of working capacity (ΔW), leading to high performance eventually. From a dynamic perspective, it was also suggested that the adsorption/desorption of working fluids majorly occurring in the temperature ranges of 305−325 and 330− 345 K was favorable for the MOFs to achieve better SCP and COP. Furthermore, the successful implementation of several commonly used machine learning (ML) algorithms paves the way for accelerating the assessment of the dynamic performance for nanoporous materials with reasonable computation time. During the training of ML algorithms, it was revealed that ΔW and transport diffusion were the dominant descriptors for predicting SCP, while equilibrium adsorption performance and MOF density played a vital role in predicting COP.

show abstract

Section: Methodsmentioning

confidence: 99%

Performance Evaluation of Metal–Organic Frameworks in Adsorption Heat Pumps via Multiscale Modeling

Liang,

Li,

et al. 2024

ACS Sustainable Chem. Eng.

View full text Add to dashboard Cite

show abstract

“…The hist gradient boosting method utilizes its algorithm to implement the processing of input variables. Every tree incorporated into an ensemble endeavors to rectify the predicted flaws by leveraging the existing models within the ensemble [ 48 ].…”

Section: Methodsmentioning

confidence: 99%

Interpretable Machine Learning Framework to Predict the Glass Transition Temperature of Polymers

Uddin,

Fan

2024

Polymers

View full text Add to dashboard Cite

The glass transition temperature of polymers is a key parameter in meeting the application requirements for energy absorption. Previous studies have provided some data from slow, expensive trial-and-error procedures. By recognizing these data, machine learning algorithms are able to extract valuable knowledge and disclose essential insights. In this study, a dataset of 7174 samples was utilized. The polymers were numerically represented using two methods: Morgan fingerprint and molecular descriptor. During preprocessing, the dataset was scaled using a standard scaler technique. We removed the features with small variance from the dataset and used the Pearson correlation technique to exclude the features that were highly connected. Then, the most significant features were selected using the recursive feature elimination method. Nine machine learning techniques were employed to predict the glass transition temperature and tune their hyperparameters. The models were compared using the performance metrics of mean absolute error (MAE), root mean square error (RMSE), and coefficient of determination (R2). We observed that the extra tree regressor provided the best results. Significant features were also identified using statistical machine learning methods. The SHAP method was also employed to demonstrate the influence of each feature on the model’s output. This framework can be adaptable to other properties at a low computational expense.

show abstract

“…Neste método, várias árvores são treinadas simultaneamente usando a técnica de boostrap ou bagging. São selecionadas diversas amostras aleatórias dos dados de treinamento, aprimorando as árvores de decisão, as quais não interagem entre si durante a fase de construção e são geradas em paralelo (Gayathri, Rani, Cepová, Rajesh, & Kalita, 2022).…”

Section: Introductionunclassified

Análise comparativa de algoritmos de machine learning para previsões de precipitação em séries observadas por estação automática em Campina Grande - PB

Neto,

Felix,

Rodrigues

et al. 2023

CLCS

View full text Add to dashboard Cite

Os recursos hídricos representam um dos mais importantes recursos naturais, devido ao seu variado escopo de finalidades. Os dados quali-quantitativos de pluviometria são essenciais para o planejamento estratégico e conhecimento pleno do modelo hidrológico. Sendo assim, gerar previsões assertivas de precipitação representam auxílio indispensável para o gerenciamento integrado de recursos, auxílio em tomadas de decisões em determinadas áreas socioeconômicas e suporte para planejamentos urbanos. Nesse sentido, esta pesquisa teve por objetivo realizar análise comparativa para a previsão em séries temporais de precipitação por meio de diferentes métodos de Machine Learning, sendo eles: Random Forest, XGBoost, Support Vector Regression e redes neurais artificiais. A pesquisa teve foco aplicado à conjunto de dados meteorológicos obtidos por estação automática em Campina Grande-PB e disponibilizado pelo Instituto Nacional de Meteorologia (INMET), no intervalo de janeiro de 2017 a dezembro de 2022. De acordo com o estudo conduzido, foram obtidos melhores resultados nas métricas de desempenho para os modelos de regressão por redes neurais multicamadas e Random Forest. Contudo, se vê necessário aprimoração na modelagem para os dados disponibilizados em análise, devido à baixa performance em todos os cenários.

show abstract

A Comparative Analysis of Machine Learning Models in Prediction of Mortar Compressive Strength

Cited by 28 publications

References 45 publications

Performance Evaluation of Metal–Organic Frameworks in Adsorption Heat Pumps via Multiscale Modeling

Performance Evaluation of Metal–Organic Frameworks in Adsorption Heat Pumps via Multiscale Modeling

Interpretable Machine Learning Framework to Predict the Glass Transition Temperature of Polymers

Análise comparativa de algoritmos de machine learning para previsões de precipitação em séries observadas por estação automática em Campina Grande - PB

Contact Info

Product

Resources

About