Crack detection based on mel-frequency cepstral coefficients features using multiple classifiers DOI Open Access
Muneera Altayeb, Areen Arabiat

International Journal of Power Electronics and Drive Systems/International Journal of Electrical and Computer Engineering, Journal Year: 2024, Volume and Issue: 14(3), P. 3332 - 3332

Published: April 4, 2024

Crack detection plays an essential role in evaluating the strength of structures. In recent years, use machine learning and deep techniques combined with computer vision has emerged to assess structures detect cracks. This research aims (ML) create a crack model based on dataset consisting 2432 images different surfaces that were divided into two groups: 70% training 30% testing dataset. The Orange3 data mining tool was used build model, where support vector (SVM), gradient boosting (GB), naive Bayes (NB), artificial neural network (ANN) trained verified 3 sets features, mel-frequency cepstral coefficients (MFCC), delta MFCC (DMFCC), delta-delta (DDMFCC) extracted using MATLAB. experimental results showed superiority SVM classification accuracy (100%), while for NB reached (93.9%-99.9%), (99.9%) ANN, finally GB (99.8%).

Language: Английский

Reliability of regression based hybrid machine learning models for the prediction of solar photovoltaics power generation DOI Creative Commons

Sina Ibne Ahmed,

Kaiser Ahmed Bhuiyan,

Irin Rahman

et al.

Energy Reports, Journal Year: 2024, Volume and Issue: 12, P. 5009 - 5023

Published: Nov. 8, 2024

Language: Английский

Citations

1
Prediction of heat transfer performance of vacuum glass based on extreme gradient boosting algorithm DOI

Feiyu Jia,

Yanggang Hu,

Lei Wang

et al.

Published: June 20, 2023

In this paper, a non-stationary detection method based on the artificial intelligence algorithm XGBoost is proposed for of U-value vacuum glass. By analyzing heat transfer characteristics glass and considering efficiency, features are selected as hot end temperature, ambient characteristic temperature change rate. training effect model measured comprehensively by scores MAE, MSE, R2. Three models, KNN, GBDT, XGBoost, used to train dataset compare prediction results. After comparison, has best effect. Finally, fitted validated 5*2 nested cross-loop, analysis results show that better stability, which greatly enhances credibility model. series experiments, it known small sample multiple interference problems can all be solved with certain provide ideas further industrialized testing.

Language: Английский

Citations

2
THE EXPLAINABILITY OF GRADIENT-BOOSTED DECISION TREES FOR DIGITAL ELEVATION MODEL (DEM) ERROR PREDICTION DOI Creative Commons
Chukwuma Okolie, J. P. Mills, Adedayo Adeleke

et al.

˜The œinternational archives of the photogrammetry, remote sensing and spatial information sciences/International archives of the photogrammetry, remote sensing and spatial information sciences, Journal Year: 2023, Volume and Issue: XLVIII-M-3-2023, P. 161 - 168

Published: Sept. 5, 2023

Abstract. Gradient boosted decision trees (GBDTs) have repeatedly outperformed several machine learning and deep algorithms in competitive data science. However, the explainability of GBDT predictions especially with earth observation is still an open issue requiring more focus by researchers. In this study, we investigate Bayesian-optimised for modelling prediction vertical error Copernicus GLO-30 digital elevation model (DEM). Three are investigated (extreme gradient boosting - XGBoost, light – LightGBM, categorical CatBoost), SHapley Additive exPlanations (SHAP) adopted analysis. The assessment sites selected from urban/industrial mountainous landscapes Cape Town, South Africa. Training datasets comprised eleven predictor variables which known influencers error: elevation, slope, aspect, surface roughness, topographic position index, terrain ruggedness texture, vector roughness measure, forest cover, bare ground urban footprints. target variable (elevation error) was calculated respect to accurate airborne LiDAR. After training testing, GBDTs were applied predicting at implementation sites. SHAP plots showed varying levels emphasis on parameters depending land cover terrain. For example, area, influence measure surpassed that first-order derivatives such as slope aspect. Thus, it recommended procedures workflows incorporate ensure robust interpretation understanding both technical non-technical users.

Language: Английский

Citations

2
Solar Radiation Prediction Using Regression Methods DOI

Saurabh Tikariha,

Yash Pal

Lecture notes in electrical engineering, Journal Year: 2024, Volume and Issue: unknown, P. 335 - 346

Published: Jan. 1, 2024

Language: Английский

Citations

0
Crack detection based on mel-frequency cepstral coefficients features using multiple classifiers DOI Open Access
Muneera Altayeb, Areen Arabiat

International Journal of Power Electronics and Drive Systems/International Journal of Electrical and Computer Engineering, Journal Year: 2024, Volume and Issue: 14(3), P. 3332 - 3332

Published: April 4, 2024

Crack detection plays an essential role in evaluating the strength of structures. In recent years, use machine learning and deep techniques combined with computer vision has emerged to assess structures detect cracks. This research aims (ML) create a crack model based on dataset consisting 2432 images different surfaces that were divided into two groups: 70% training 30% testing dataset. The Orange3 data mining tool was used build model, where support vector (SVM), gradient boosting (GB), naive Bayes (NB), artificial neural network (ANN) trained verified 3 sets features, mel-frequency cepstral coefficients (MFCC), delta MFCC (DMFCC), delta-delta (DDMFCC) extracted using MATLAB. experimental results showed superiority SVM classification accuracy (100%), while for NB reached (93.9%-99.9%), (99.9%) ANN, finally GB (99.8%).

Language: Английский

Citations

0