Comparative Study of Hybrid Artificial Intelligence Approaches for Predicting Hangingwall Stability

Qi, Chongchong; Fourie, Andy; Ma, Guowei; Tang, Xiaolin; Du, Xuhao

doi:10.1061/(asce)cp.1943-5487.0000737

Cited by 58 publications

(28 citation statements)

References 34 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Figure also shows that the dataset was quite widely distributed and the distribution of most variables was not symmetric. Therefore, normalising all variables into [0, 1] range based on their corresponding minimum and maximum values may improve the computation efficiency of classifiers …”

Section: Methodsmentioning

confidence: 99%

“…The training set is used to build classifiers, while the testing set is used to verify their predictive performance. In AI, the percentage of the training set to the whole dataset needs to be determined by an optimisation analysis to obtain a representative training set . In this paper, 70% of the whole dataset (58 unstable slopes and 59 stable slopes) was included in the training set, and the remaining 30% (26 unstable slopes and 25 stable slopes) was included in the testing set after the optimisation analysis.…”

Section: Methodsmentioning

confidence: 99%

“…A series of points will be connected so that a ROC curve could be generated. Several methods have been proposed for the selection of cut‐offs, such as the baseline cut‐off (0.5), the Youden's cut‐off (cut‐off that achieves the largest Youden's J index), and the top left cut‐off (cut‐off that is closest to the top left point in the ROC curve) . In this paper, the baseline cut‐off and the Youden's cut‐off were utilised for the cut‐off selection in slope stability analysis.…”

Section: Methodsmentioning

confidence: 99%

“…In this section, six individual AI techniques, including GPC, QDA, SVM, ANN, ADB‐DT, and the classifier ensembles, are introduced. The selection of these individual techniques was based on their shared characteristics: All techniques are good at modelling the nonlinear relationships between inputs and outputs, which are required in the prediction of slope stability All techniques are being increasingly used in related civil engineering problems and showed good predictive performance in related works …”

Section: Individual Ai Techniques and The Classifier Ensemblesmentioning

confidence: 99%

“…For limit equilibrium methods, the search for the critical slip surface is cumbersome as there are a large number of potential slip surfaces . For numerical modelling, the determination of constitutive models and mechanical parameters, which often requires rich engineering experience and in situ back analysis, is difficult . Therefore, there still exist considerable challenges in understanding and predicting slope stability.…”

Section: Introductionmentioning

confidence: 99%

See 4 more Smart Citations

A hybrid ensemble method for improved prediction of slope stability

Tang

2018

Num Anal Meth Geomechanics

Self Cite

View full text Add to dashboard Cite

Accurate prediction of slope stability is a significant issue in geomechanics with many artificial intelligence (AI) techniques being utilised. However, the application of AI has not reached its full potential because of the lack of more robust algorithms. In this paper, we proposed a hybrid ensemble method for the improved prediction of slope stability using classifier ensembles and genetic algorithm. Gaussian process classification, quadratic discriminant analysis, support vector machine, artificial neural networks, adaptive boosted decision trees, and k-nearest neighbours were chosen to be individual AI techniques, and the weighted majority voting was used as the combination method. Validation method was chosen to be the 10-fold cross-validation, and performance measures were selected to be the accuracy, the receiver operating characteristic curve, and the area under the receiver operating characteristic curve (AUC). Grid search and genetic algorithm were used for the hyperparameter tuning and weight tuning respectively. The results show that the proposed hybrid ensemble method has great potential in improving the prediction of slope stability. Compared with individual classifiers, the optimum ensemble classifier achieved the highest AUC value (0.943) and the highest accuracy (0.902) on the testing set, denoting that the predictive performance has been improved. The optimum ensemble classifier with the Youden's cutoff was recommended for slope stability prediction with respect to the AUC value, the accuracy, the true positive rate, and the true negative rate. This research indicates that the use of the classifier ensembles, rather than the search for the ideal individual classifiers, might help for the slope stability prediction.

show abstract

Section: Methodsmentioning

confidence: 99%

Section: Methodsmentioning

confidence: 99%

Section: Methodsmentioning

confidence: 99%

Section: Individual Ai Techniques and The Classifier Ensemblesmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

A hybrid ensemble method for improved prediction of slope stability

Tang

2018

Num Anal Meth Geomechanics

Self Cite

View full text Add to dashboard Cite

show abstract

Convolutional neural network-based data anomaly detection method using multiple information for structural health monitoring

Tang

Chen

Bao

et al. 2018

Struct Control Health Monit

277

156

View full text Add to dashboard Cite

Structural health monitoring (SHM) is used worldwide for managing and maintaining civil infrastructures. SHM systems have produced huge amounts of data, but the effective monitoring, mining, and utilization of this data still need in-depth study. SHM data generally includes multiple types of anomalies caused by sensor faults or system malfunctions that can disturb structural analysis and assessment. In the routine data pre-processing, multiple signal processing techniques are required to detect the anomalies, respectively, which is inefficient. The large variations of extracted features from massive SHM data make the data anomaly detection techniques prone to be over-processed or under-processed. Even with expert intervention, the parameter tuning, associated with multiple data preprocessing methods, is still a challenge, making the procedure expensive and inefficient. In addition, one data anomaly detection technique frequently mis-detects other types of anomaly. In this work, we focus on the anomaly detection in the stage of data pre-processing that little work has been done based on the real-world continuous SHM data with multiclass anomalies. We proposed a novel data anomaly detection method based on a convolutional neural network (CNN) that imitates human vision and decision making. First, we split raw time series data into sections, and visualized the data in time and frequency domain, respectively. Then each section's images were stacked as a single dual-channel image and labeled according to graphical features (multi-2D image space expression). Second, a CNN was designed and trained for data anomaly classification, during which the descriptions and representations of the anomalies' features were learned by convolution. To validate our work, we considered the effects of balanced and imbalanced training sets and training ratios on actual acceleration data from an SHM system for a long-span cable-stayed bridge. The results show that our method could detect the multipattern anomalies of SHM data efficiently with high accuracy. The proposed dual-information CNN-based design makes this detection process readily scalable, faster, and more accurate, thereby providing a novel perspective with strong potential for SHM data preprocessing. KEYWORDScomputer vision, convolutional neural network (CNN), data anomaly detection, long-span bridge, structural health monitoring (SHM) | INTRODUCTIONStructural health monitoring (SHM) is used worldwide to manage and maintain civil infrastructure systems by evaluating their structural loads, responses, and real-time performance, and by predicting the future behavior of structures of all types. 1-4 SHM has produced huge amounts of data. For example, the SHM system used for the Sutong Bridge in China, which has 785 sensors, produces 2.5 TB of data each year. The effective mining and utilization of SHM data is an important topic that needs in-depth study. However, researchers have faced challenges because the SHM data generally include multiple types of anomalies caused by sensor fa...

show abstract