Failure mode and effects analysis of RC members based on machine-learning-based SHapley Additive exPlanations (SHAP) approach

Mangalathu, Sujith; Hwang, Seong-Hoon; Jeon, Jong‐Su

doi:10.1016/j.engstruct.2020.110927

Cited by 466 publications

(161 citation statements)

References 20 publications

Supporting

Mentioning

123

Contrasting

Order By: Relevance

“…Several papers have been published on the applications of artificial neural networks (ANNs) for buckling analysis of beam-columns [28] and cellular steel beams [29], as well as for different aspects of CFS structures, including space structure optimization [30], web crippling strength prediction [31], elastic distortional buckling stress determination [32,33], and rotation capacity prediction [34]. The majority of ML algorithms considered in this work were employed for structural concrete: support vector machines (SVM) were studied in [35][36][37][38][39][40][41], decision trees (DT) in [40,42], random forest (RF) in [40,[42][43][44][45], and k-nearest neighbors (KNN) in [37,40,42,46]. Steel and steel-concrete composite structures have also seen limited applications of SVM [47][48][49] and DT and KNN [50].…”

Section: Introductionmentioning

confidence: 99%

Predicting shear strength of CFS channels with slotted webs by machine learning models

Degtyarev

2021

Archit. Struct. Constr.

View full text Add to dashboard Cite

Staggered rectangular perforations (slots) are provided in the webs of cold-formed steel (CFS) beams and columns to reduce their thermal conductivity and improve the energy efficiency of CFS buildings. The perforations adversely affect the structural characteristics of the members, especially those governed by the web parameters, such as the shear strength and shear buckling. This paper presents machine learning (ML) models to predict the elastic shear buckling load and the ultimate shear strength of CFS channels with slotted webs. Support vector machine (SVM), decision tree (DT), random forest (RF), and k-nearest neighbor (KNN) regressors were trained using a large dataset of numerical results with 3512 samples. An extensive search was conducted to find optimal hyperparameters of the models that result in the best predictions and prevent overfitting. The models' performances were evaluated employing the ten-fold cross-validation method to make more data available for training and reduce bias and variance. The SVR, DT, and RF models demonstrate good prediction accuracy, which exceeds the accuracy of the existing descriptive equations. Relative feature importance was evaluated using the permutation and SHAP methods for each model. Partial dependence of the buckling load and the shear strength from the channel features was assessed. The predictions of the developed ML models were also compared with the predictions of previously developed artificial neural networks (ANNs). The comparisons demonstrated that ANNs showed higher accuracy in predicting the elastic buckling load, whereas the SVR model provided the most accurate shear strength predictions. Keywords Cold-formed steel (CFS) • Perforated web • Shear buckling • Shear capacity • Machine learning • Predictive models Publisher's Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

show abstract

Section: Introductionmentioning

confidence: 99%

Predicting shear strength of CFS channels with slotted webs by machine learning models

Degtyarev

2021

Archit. Struct. Constr.

View full text Add to dashboard Cite

show abstract

“…By allocating a significance value (SHAP value) to any variable that meets the following criteria, it explains a particular prediction using Shapley values. The requirements [ 54 ] include (1) consistency—if we change a model to make it more dependent on a particular component, the relevance of that component should not diminish, notwithstanding the relevance of other variables; (2) missingness—components that aren’t present in the primary input must be ignored; (3) local accuracy—the explanation method has to at least in accordance with the main model’s output. Therefore, SHAP can describe both global and local methods effectively.…”

Section: Methodsmentioning

confidence: 99%

Section: Shap Approachmentioning

confidence: 99%

Urban Vegetation Mapping from Aerial Imagery Using Explainable AI (XAI)

Abdollahi

Pradhan

2021

Sensors

View full text Add to dashboard Cite

Urban vegetation mapping is critical in many applications, i.e., preserving biodiversity, maintaining ecological balance, and minimizing the urban heat island effect. It is still challenging to extract accurate vegetation covers from aerial imagery using traditional classification approaches, because urban vegetation categories have complex spatial structures and similar spectral properties. Deep neural networks (DNNs) have shown a significant improvement in remote sensing image classification outcomes during the last few years. These methods are promising in this domain, yet unreliable for various reasons, such as the use of irrelevant descriptor features in the building of the models and lack of quality in the labeled image. Explainable AI (XAI) can help us gain insight into these limits and, as a result, adjust the training dataset and model as needed. Thus, in this work, we explain how an explanation model called Shapley additive explanations (SHAP) can be utilized for interpreting the output of the DNN model that is designed for classifying vegetation covers. We want to not only produce high-quality vegetation maps, but also rank the input parameters and select appropriate features for classification. Therefore, we test our method on vegetation mapping from aerial imagery based on spectral and textural features. Texture features can help overcome the limitations of poor spectral resolution in aerial imagery for vegetation mapping. The model was capable of obtaining an overall accuracy (OA) of 94.44% for vegetation cover mapping. The conclusions derived from SHAP plots demonstrate the high contribution of features, such as Hue, Brightness, GLCM_Dissimilarity, GLCM_Homogeneity, and GLCM_Mean to the output of the proposed model for vegetation mapping. Therefore, the study indicates that existing vegetation mapping strategies based only on spectral characteristics are insufficient to appropriately classify vegetation covers.

show abstract

“…In turn, SHAP [36], [37] is qualitatively different from the simple search for correlations, in view of the fact that it uses the model for gaining knowledge about nonlinear and nonmonotonic interdependencies of parameters that influence the final result. SHAP is widely used to explain the result of machine learning [38]- [42].…”

Section: Accounting Methods Of Expert Assessments Inconsistencymentioning

confidence: 99%

Assessing the Impact of Expert Labelling of Training Data on the Quality of Automatic Classification of Lithological Groups Using Artificial Neural Networks

Kuchin

Mukhamediev

Yakunin

et al. 2020

Applied Computer Systems

View full text Add to dashboard Cite

Machine learning (ML) methods are nowadays widely used to automate geophysical study. Some of ML algorithms are used to solve lithological classification problems during uranium mining process. One of the key aspects of using classical ML methods is causing data features and estimating their influence on the classification. This paper presents a quantitative assessment of the impact of expert opinions on the classification process. In other words, we have prepared the data, identified the experts and performed a series of experiments with and without taking into account the fact that the expert identifier is supplied to the input of the automatic classifier during training and testing. Feedforward artificial neural network (ANN) has been used as a classifier. The results of the experiments show that the “knowledge” of the ANN of which expert interpreted the data improves the quality of the automatic classification in terms of accuracy (by 5 %) and recall (by 20 %). However, due to the fact that the input parameters of the model may depend on each other, the SHapley Additive exPlanations (SHAP) method has been used to further assess the impact of expert identifier. SHAP has allowed assessing the degree of parameter influence. It has revealed that the expert ID is at least two times more influential than any of the other input parameters of the neural network. This circumstance imposes significant restrictions on the application of ANNs to solve the task of lithological classification at the uranium deposits.

show abstract

Failure mode and effects analysis of RC members based on machine-learning-based SHapley Additive exPlanations (SHAP) approach

Cited by 466 publications

References 20 publications

Predicting shear strength of CFS channels with slotted webs by machine learning models

Predicting shear strength of CFS channels with slotted webs by machine learning models

Urban Vegetation Mapping from Aerial Imagery Using Explainable AI (XAI)

Assessing the Impact of Expert Labelling of Training Data on the Quality of Automatic Classification of Lithological Groups Using Artificial Neural Networks

Contact Info

Product

Resources

About