Machine learning has been widely used for solving partial differential equations(PDEs)in recent years,among which the random feature method(RFM)exhibits spectral accuracy and can compete with traditional solvers in te...Machine learning has been widely used for solving partial differential equations(PDEs)in recent years,among which the random feature method(RFM)exhibits spectral accuracy and can compete with traditional solvers in terms of both accuracy and efficiency.Potentially,the optimization problem in the RFM is more difficult to solve than those that arise in traditional methods.Unlike the broader machine-learning research,which frequently targets tasks within the low-precision regime,our study focuses on the high-precision regime crucial for solving PDEs.In this work,we study this problem from the following aspects:(i)we analyze the coeffcient matrix that arises in the RFM by studying the distribution of singular values;(ii)we investigate whether the continuous training causes the overfitting issue;(ii)we test direct and iterative methods as well as randomized methods for solving the optimization problem.Based on these results,we find that direct methods are superior to other methods if memory is not an issue,while iterative methods typically have low accuracy and can be improved by preconditioning to some extent.展开更多
Piled embankments are widely used in highway and railway engineering due to their economy and efficiency inovercoming several issues encountered in constructing embankments over weak soils. Soil arching, caused by the...Piled embankments are widely used in highway and railway engineering due to their economy and efficiency inovercoming several issues encountered in constructing embankments over weak soils. Soil arching, caused by the pile-subsoilrelative displacement （△s）, plays an important role in reducing the embankment load falling on weak soil, however, the funda-mental characteristics （e.g., formation and features） of soil arching remain poorly understood. In this study, a series of discreteelement method （DEM） modellings are performed to study the formation and features of soil arching with the variation of As inpiled embankments with or without geosynthetic reinforcement. Firstly, calibration for the modelling parameters is carried out bycomparing the DEM results with the experimental data obtained from the existing literature. Secondly, the analysis of the macro-and micro-behaviours is performed in detail. Finally, a parametric study is conducted in an effort to identify the influences of threekey factors on soil arching： the friction coefficient of the embankment fill （f）, the embankment height （h）, and the pile clear spacing（s-a）. Numerical results indicate that △s is a key factor governing the formation and features of soil arching in embankments. Tobe specific, soil arching gradually evolves from two inclined shear planes at a small △s to a hemispherical arch at a relatively largeAs. Then, with a continuous increase in △s, the soil arching height gradually increases and finally approaches a constant value of0.8（s-a） （i.e., the maximum soil arching height）. For a given case, the higher the soil arching height, the greater the degree of soilarching effect. The parametric study shows that the friction coefficient of the embankment fill has a negligible influence on theformation and features of soil arching. However, embankment height is a key factor governing the formation and features of soilarching. In addition, pile clear spacing has a significant effect on the formation of soil arching, but not on its features.展开更多
Aiming at the existing problems of discrete cosine transform(DCT) de-noising method, we introduce the idea of wavelet neighboring coefficients(WNC) de-noising method, and propose the cosine neighboring coefficients(CN...Aiming at the existing problems of discrete cosine transform(DCT) de-noising method, we introduce the idea of wavelet neighboring coefficients(WNC) de-noising method, and propose the cosine neighboring coefficients(CNC) de-noising method. Based on DCT, a novel method for the fault feature extraction of hydraulic pump is analyzed. The vibration signal of pump is de-noised with CNC de-noising method, and the fault feature is extracted by performing Hilbert-Huang transform(HHT) to the output signal. The analysis results of the simulation signal and the actual one demonstrate that the proposed CNC de-noising method and the fault feature extraction method have more superior ability than the traditional ones.展开更多
The main purpose of reverse engineering is to convert discrete data pointsinto piecewise smooth, continuous surface models. Before carrying out model reconstruction it issignificant to extract geometric features becau...The main purpose of reverse engineering is to convert discrete data pointsinto piecewise smooth, continuous surface models. Before carrying out model reconstruction it issignificant to extract geometric features because the quality of modeling greatly depends on therepresentation of features. Some fitting techniques of natural quadric surfaces with least-squaresmethod are described. And these techniques can be directly used to extract quadric surfaces featuresduring the process of segmentation for point cloud.展开更多
Accurate prediction of shield tunneling-induced settlement is a complex problem that requires consideration of many influential parameters.Recent studies reveal that machine learning(ML)algorithms can predict the sett...Accurate prediction of shield tunneling-induced settlement is a complex problem that requires consideration of many influential parameters.Recent studies reveal that machine learning(ML)algorithms can predict the settlement caused by tunneling.However,well-performing ML models are usually less interpretable.Irrelevant input features decrease the performance and interpretability of an ML model.Nonetheless,feature selection,a critical step in the ML pipeline,is usually ignored in most studies that focused on predicting tunneling-induced settlement.This study applies four techniques,i.e.Pearson correlation method,sequential forward selection(SFS),sequential backward selection(SBS)and Boruta algorithm,to investigate the effect of feature selection on the model’s performance when predicting the tunneling-induced maximum surface settlement(S_(max)).The data set used in this study was compiled from two metro tunnel projects excavated in Hangzhou,China using earth pressure balance(EPB)shields and consists of 14 input features and a single output(i.e.S_(max)).The ML model that is trained on features selected from the Boruta algorithm demonstrates the best performance in both the training and testing phases.The relevant features chosen from the Boruta algorithm further indicate that tunneling-induced settlement is affected by parameters related to tunnel geometry,geological conditions and shield operation.The recently proposed Shapley additive explanations(SHAP)method explores how the input features contribute to the output of a complex ML model.It is observed that the larger settlements are induced during shield tunneling in silty clay.Moreover,the SHAP analysis reveals that the low magnitudes of face pressure at the top of the shield increase the model’s output。展开更多
Aiming to deficiency of the filter and wrapper feature selection methods, anew method based on composite method of filter and wrapper method is proposed. First the methodfilters original features to form a feature sub...Aiming to deficiency of the filter and wrapper feature selection methods, anew method based on composite method of filter and wrapper method is proposed. First the methodfilters original features to form a feature subset which can meet classification correctness rate,then applies wrapper feature selection method select optimal feature subset. A successful techniquefor solving optimization problems is given by genetic algorithm (GA). GA is applied to the problemof optimal feature selection. The composite method saves computing time several times of the wrappermethod with holding the classification accuracy in data simulation and experiment on bearing faultfeature selection. So this method possesses excellent optimization property, can save more selectiontime, and has the characteristics of high accuracy and high efficiency.展开更多
An efficient multi-harmonic method is proposed for studying the effects of mistuning on resonant features of bladed disks with blade-to-blade dry friction damping. This method is able to predict accurately the forced ...An efficient multi-harmonic method is proposed for studying the effects of mistuning on resonant features of bladed disks with blade-to-blade dry friction damping. This method is able to predict accurately the forced response of bladed disks in frequency domain, which is validated by numerical integration method in time domain. The resonant features of both tuned and mistuned systems are investigated by using this method under various system coupling strengths, viscous dampings, and dry friction darnpings, etc. The results demonstrate that the proposed multi-harmonic method is very efficient for studying the mistuning effects on the resonant response of bladed disks with blade-to-blade dry friction damping, especially considering the combined effects of various system parameters.展开更多
In order to provide a guidance to specify the element size dynamically during adaptive finite element mesh generation, adaptive criteria are firstly defined according to the relationships between the geometrical featu...In order to provide a guidance to specify the element size dynamically during adaptive finite element mesh generation, adaptive criteria are firstly defined according to the relationships between the geometrical features and the elements of 3D solid. Various modes based on different datum geometrical elements, such as vertex, curve, surface, and so on, are then designed for generating local refined mesh. With the guidance of the defmed criteria, different modes are automatically selected to apply on the appropriate datum objects to program the element size in the local special areas. As a result, the control information of element size is successfully programmed covering the entire domain based on the geometrical features of 3D solid. A new algorithm based on Delatmay triangulation is then developed for generating 3D adaptive finite element mesh, in which the element size is dynamically specified to catch the geometrical features and suitable tetrahedron facets are selected to locate interior nodes continuously. As a result, adaptive mesh with good-quality elements is generated. Examples show that the proposed method can be successfully applied to adaptive finite element mesh automatic generation based on the geometrical features of 3D solid.展开更多
Several millions of people suffer from Parkinson’s disease globally.Parkinson’s affects about 1%of people over 60 and its symptoms increase with age.The voice may be affected and patients experience abnormalities in...Several millions of people suffer from Parkinson’s disease globally.Parkinson’s affects about 1%of people over 60 and its symptoms increase with age.The voice may be affected and patients experience abnormalities in speech that might not be noticed by listeners,but which could be analyzed using recorded speech signals.With the huge advancements of technology,the medical data has increased dramatically,and therefore,there is a need to apply data mining and machine learning methods to extract new knowledge from this data.Several classification methods were used to analyze medical data sets and diagnostic problems,such as Parkinson’s Disease(PD).In addition,to improve the performance of classification,feature selection methods have been extensively used in many fields.This paper aims to propose a comprehensive approach to enhance the prediction of PD using several machine learning methods with different feature selection methods such as filter-based and wrapper-based.The dataset includes 240 recodes with 46 acoustic features extracted from3 voice recording replications for 80 patients.The experimental results showed improvements when wrapper-based features selection method was used with K-NN classifier with accuracy of 88.33%.The best obtained results were compared with other studies and it was found that this study provides comparable and superior results.展开更多
The development of defect prediction plays a significant role in improving software quality. Such predictions are used to identify defective modules before the testing and to minimize the time and cost. The software w...The development of defect prediction plays a significant role in improving software quality. Such predictions are used to identify defective modules before the testing and to minimize the time and cost. The software with defects negatively impacts operational costs and finally affects customer satisfaction. Numerous approaches exist to predict software defects. However, the timely and accurate software bugs are the major challenging issues. To improve the timely and accurate software defect prediction, a novel technique called Nonparametric Statistical feature scaled QuAdratic regressive convolution Deep nEural Network (SQADEN) is introduced. The proposed SQADEN technique mainly includes two major processes namely metric or feature selection and classification. First, the SQADEN uses the nonparametric statistical Torgerson–Gower scaling technique for identifying the relevant software metrics by measuring the similarity using the dice coefficient. The feature selection process is used to minimize the time complexity of software fault prediction. With the selected metrics, software fault perdition with the help of the Quadratic Censored regressive convolution deep neural network-based classification. The deep learning classifier analyzes the training and testing samples using the contingency correlation coefficient. The softstep activation function is used to provide the final fault prediction results. To minimize the error, the Nelder–Mead method is applied to solve non-linear least-squares problems. Finally, accurate classification results with a minimum error are obtained at the output layer. Experimental evaluation is carried out with different quantitative metrics such as accuracy, precision, recall, F-measure, and time complexity. The analyzed results demonstrate the superior performance of our proposed SQADEN technique with maximum accuracy, sensitivity and specificity by 3%, 3%, 2% and 3% and minimum time and space by 13% and 15% when compared with the two state-of-the-art methods.展开更多
Mahalanobis-Taguchi system(MTS)is a kind of data mining and pattern recognition method which can identify the attribute characteristics of multidimensional data by constructing Mahalanobis distance(MD)measurement scal...Mahalanobis-Taguchi system(MTS)is a kind of data mining and pattern recognition method which can identify the attribute characteristics of multidimensional data by constructing Mahalanobis distance(MD)measurement scale.In this paper,considering the influence of irregular distribution of the sample data and abnormal variation of the normal data on accuracy of MTS,a feature recognition and selection model of the equipment state based on the improved MTS is proposed,and two aspects of the model namely construction of the original Mahalanobis space(MS)and determination of the threshold are studied.Firstly,the original training sample space is statistically controlled by the X-bar-S control chart,and extreme data of the single characteristic attribute is filtered to reduce the impact of extreme condition on the accuracy of the model,so as to construct a more robust MS.Furthermore,the box plot method is used to determine the threshold of the model.And the stability of the model and the tolerance to the extreme condition are improved by leaving sufficient range of the variation for the extreme condition which is identified as in the normal range.Finally,the improved model is compared with the traditional one based on the unimproved MTS by using the data from the literature.The result shows that compared with the traditional model,the accuracy and sensitivity of the improved model for state identification can be greatly enhanced.展开更多
This paper presents a multiscale design method for simultaneous topology optimization of both macrostructures and microstructures.Geometric features are extended as design primitives at both macro and micro scales and...This paper presents a multiscale design method for simultaneous topology optimization of both macrostructures and microstructures.Geometric features are extended as design primitives at both macro and micro scales and represented by Level Set Functions(LSFs).Parameters related to the locations,sizes,and orientations of macro and micro features are considered as design variables and optimized simultaneously.In the overlapping areas of different macro features,embedded microstructures are optimally figured out as the solution of the corresponding sub-optimization,problem.In this study,the eXtended Finite Element Method(XFEM)is implemented for structural and sensitivity analyses with respect to design variables.This method has the advantage of using a fixed grid independent of the topology optimization process.The homogenization procedure is applied to calculate the effective properties of considered microstructures in each macro feature.Numerical examples are presented to illustrate the effectiveness of the proposed method.Results depict that the multiscale design cannot obviously improve structural stiffness compared with a solid-material design under the linear elastic condition.展开更多
The feature-selection problem in training AdaBoost classifiers is addressed in this paper. A working feature subset is generated by adopting a novel feature subset selection method based on the partial least square （...The feature-selection problem in training AdaBoost classifiers is addressed in this paper. A working feature subset is generated by adopting a novel feature subset selection method based on the partial least square （PLS） regression, and then trained and selected from this feature subset in Boosting. The experiments show that the proposed PLS-based feature-selection method outperforms the current feature ranking method and the random sampling method.展开更多
We describe how the Unit-Feature Spatial Classification Method(UFSCM) can be used operationally to classify cloud types in satellite imagery efficiently and conveniently.By using a combination of Interactive Data Lang...We describe how the Unit-Feature Spatial Classification Method(UFSCM) can be used operationally to classify cloud types in satellite imagery efficiently and conveniently.By using a combination of Interactive Data Language(IDL) and Visual C++(VC) code in combination to extend the technique in three dimensions(3-D),this paper provides an efficient method to implement interactive computer visualization of the 3-D discrimination matrix modification,so as to deal with the bi-spectral limitations of traditional two dimensional(2-D) UFSCM.The case study of cloud-type classification based on FY-2C satellite data (0600 UTC 18 and 0000 UTC 10 September 2007) is conducted by comparison with ground station data, and indicates that 3-D UFSCM makes more use of the pattern recognition information in multi-spectral imagery,resulting in more reasonable results and an improvement over the 2-D method.展开更多
In order to improve the accuracy of building structure identification using remote sensing images,a building structure classification method based on multi-feature fusion of UAV remote sensing image is proposed in thi...In order to improve the accuracy of building structure identification using remote sensing images,a building structure classification method based on multi-feature fusion of UAV remote sensing image is proposed in this paper.Three identification approaches of remote sensing images are integrated in this method:object-oriented,texture feature,and digital elevation based on DSM and DEM.So RGB threshold classification method is used to classify the identification results.The accuracy of building structure classification based on each feature and the multi-feature fusion are compared and analyzed.The results show that the building structure classification method is feasible and can accurately identify the structures in large-area remote sensing images.展开更多
ICSED (Improved Cluster Shade Edge Detection) algorithm and other various methods to accurately and efficiently detect edges on satellite data are presented. Error rate criterion is used to statistically evaluate the ...ICSED (Improved Cluster Shade Edge Detection) algorithm and other various methods to accurately and efficiently detect edges on satellite data are presented. Error rate criterion is used to statistically evaluate the performances of these methods in detecting oceanic features for both noise free and noise contaminated AVHRR (Advanced Very High Resolution Radiometer) IR image with Kuroshio. Also, practical experiments in detecting the eddy of Kuroshio with these methods are carried out for comparison. Results show that the ICSED algorithm has more advantages than other methods in detecting mesoscale features of ocean. Finally, the effectiveness of window size of ICSED method to oceanic features detection is quantitatively discussed.展开更多
Web data extraction has become a key technology for extracting valuable data from websites.At present,most extraction methods based on rule learning,visual pattern or tree matching have limited performance on complex ...Web data extraction has become a key technology for extracting valuable data from websites.At present,most extraction methods based on rule learning,visual pattern or tree matching have limited performance on complex web pages.Through ana-lyzing various statistical characteristics of HTML el-ements in web documents,this paper proposes,based on statistical features,an unsupervised web data ex-traction method—traversing the HTML DOM parse tree at first,calculating and generating the statistical matrix of the elements,and then locating data records by clustering method and heuristic rules that reveal in-herent links between the visual characteristics of the data recording areas and the statistical characteristics of the HTML nodes—which is both suitable for data records extraction of single-page and multi-pages,and it has strong generality and needs no training.The ex-periments show that the accuracy and efficiency of this method are equally better than the current data extrac-tion method.展开更多
Electrical resistivity tomography (ERT) has been used to experimentally detect shallow buried faults in urban areas in the past a few years, with some progress and experience obtained. According to the results from Ol...Electrical resistivity tomography (ERT) has been used to experimentally detect shallow buried faults in urban areas in the past a few years, with some progress and experience obtained. According to the results from Olympic Park, Beijing, Shandong Province, Gansu Province and Shanxi Province, we have generalized the method and procedure for inferring the discontinuity of electrical structures (DES) indicating a buried fault in urban areas from resistivity tomograms and its typical electrical features. In general, the layered feature of the electrical structure is first analyzed to preliminarily define whether or not a DES exists in the target area. Resistivity contours in resistivity tomograms are then analyzed from the deep to the shallow. If they extend upward from the deep to the shallow and shape into an integral dislocation, sharp flexure (convergence) or gradient zone, it is inferred that the DES exists, indicating a buried fault. Finally, horizontal tracing is be carried out to define the trend of the DES. The DES can be divided into three types-type AB, ABA and AC. In the present paper, the Zhangdian-Renhe fault system in Zibo city is used as an example to illustrate how to use the method to infer the location and spatial extension of a target fault. Geologic drilling holes are placed based on our research results, and the drilling logs testify that our results are correct. However, the method of this paper is not exclusive and inflexible. It is expected to provide reference and assistance for inferring the shallow buried faults in urban areas from resistivity tomograms in the future.展开更多
We propose a hierarchical multi-scale attention mechanism-based model in response to the low accuracy and inefficient manual classification of existing oceanic biological image classification methods. Firstly, the hie...We propose a hierarchical multi-scale attention mechanism-based model in response to the low accuracy and inefficient manual classification of existing oceanic biological image classification methods. Firstly, the hierarchical efficient multi-scale attention(H-EMA) module is designed for lightweight feature extraction, achieving outstanding performance at a relatively low cost. Secondly, an improved EfficientNetV2 block is used to integrate information from different scales better and enhance inter-layer message passing. Furthermore, introducing the convolutional block attention module(CBAM) enhances the model's perception of critical features, optimizing its generalization ability. Lastly, Focal Loss is introduced to adjust the weights of complex samples to address the issue of imbalanced categories in the dataset, further improving the model's performance. The model achieved 96.11% accuracy on the intertidal marine organism dataset of Nanji Islands and 84.78% accuracy on the CIFAR-100 dataset, demonstrating its strong generalization ability to meet the demands of oceanic biological image classification.展开更多
Hyperspectral image provides abundant spectral information for remote discrimination of subtle differences in ground covers.However,the increasing spectral dimensions,as well as the information redundancy,make the ana...Hyperspectral image provides abundant spectral information for remote discrimination of subtle differences in ground covers.However,the increasing spectral dimensions,as well as the information redundancy,make the analysis and interpretation of hyperspectral images a challenge.Feature extraction is a very important step for hyperspectral image processing.Feature extraction methods aim at reducing the dimension of data,while preserving as much information as possible.Particularly,nonlinear feature extraction methods (e.g.kernel minimum noise fraction (KMNF) transformation) have been reported to benefit many applications of hyperspectral remote sensing,due to their good preservation of high-order structures of the original data.However,conventional KMNF or its extensions have some limitations on noise fraction estimation during the feature extraction,and this leads to poor performances for post-applications.This paper proposes a novel nonlinear feature extraction method for hyperspectral images.Instead of estimating noise fraction by the nearest neighborhood information (within a sliding window),the proposed method explores the use of image segmentation.The approach benefits both noise fraction estimation and information preservation,and enables a significant improvement for classification.Experimental results on two real hyperspectral images demonstrate the efficiency of the proposed method.Compared to conventional KMNF,the improvements of the method on two hyperspectral image classification are 8 and 11％.This nonlinear feature extraction method can be also applied to other disciplines where high-dimensional data analysis is required.展开更多
基金supported by the NSFC Major Research Plan--Interpretable and Generalpurpose Next-generation Artificial Intelligence(No.92370205).
文摘Machine learning has been widely used for solving partial differential equations(PDEs)in recent years,among which the random feature method(RFM)exhibits spectral accuracy and can compete with traditional solvers in terms of both accuracy and efficiency.Potentially,the optimization problem in the RFM is more difficult to solve than those that arise in traditional methods.Unlike the broader machine-learning research,which frequently targets tasks within the low-precision regime,our study focuses on the high-precision regime crucial for solving PDEs.In this work,we study this problem from the following aspects:(i)we analyze the coeffcient matrix that arises in the RFM by studying the distribution of singular values;(ii)we investigate whether the continuous training causes the overfitting issue;(ii)we test direct and iterative methods as well as randomized methods for solving the optimization problem.Based on these results,we find that direct methods are superior to other methods if memory is not an issue,while iterative methods typically have low accuracy and can be improved by preconditioning to some extent.
基金supported by the National Key Research and Development Program of China(2016YFC0800208)the National Natural Science Foundation of China(Nos.51278216,51478201,51308241,and 51608316)
文摘Piled embankments are widely used in highway and railway engineering due to their economy and efficiency inovercoming several issues encountered in constructing embankments over weak soils. Soil arching, caused by the pile-subsoilrelative displacement （△s）, plays an important role in reducing the embankment load falling on weak soil, however, the funda-mental characteristics （e.g., formation and features） of soil arching remain poorly understood. In this study, a series of discreteelement method （DEM） modellings are performed to study the formation and features of soil arching with the variation of As inpiled embankments with or without geosynthetic reinforcement. Firstly, calibration for the modelling parameters is carried out bycomparing the DEM results with the experimental data obtained from the existing literature. Secondly, the analysis of the macro-and micro-behaviours is performed in detail. Finally, a parametric study is conducted in an effort to identify the influences of threekey factors on soil arching： the friction coefficient of the embankment fill （f）, the embankment height （h）, and the pile clear spacing（s-a）. Numerical results indicate that △s is a key factor governing the formation and features of soil arching in embankments. Tobe specific, soil arching gradually evolves from two inclined shear planes at a small △s to a hemispherical arch at a relatively largeAs. Then, with a continuous increase in △s, the soil arching height gradually increases and finally approaches a constant value of0.8（s-a） （i.e., the maximum soil arching height）. For a given case, the higher the soil arching height, the greater the degree of soilarching effect. The parametric study shows that the friction coefficient of the embankment fill has a negligible influence on theformation and features of soil arching. However, embankment height is a key factor governing the formation and features of soilarching. In addition, pile clear spacing has a significant effect on the formation of soil arching, but not on its features.
基金the National Natural Science Foundation of China(No.51275524)the General Armaments Department Equipment Support Research Project
文摘Aiming at the existing problems of discrete cosine transform(DCT) de-noising method, we introduce the idea of wavelet neighboring coefficients(WNC) de-noising method, and propose the cosine neighboring coefficients(CNC) de-noising method. Based on DCT, a novel method for the fault feature extraction of hydraulic pump is analyzed. The vibration signal of pump is de-noised with CNC de-noising method, and the fault feature is extracted by performing Hilbert-Huang transform(HHT) to the output signal. The analysis results of the simulation signal and the actual one demonstrate that the proposed CNC de-noising method and the fault feature extraction method have more superior ability than the traditional ones.
基金This project is supported by Research Foundation for Doctoral Program of Higher Education, China (No.98033532)
文摘The main purpose of reverse engineering is to convert discrete data pointsinto piecewise smooth, continuous surface models. Before carrying out model reconstruction it issignificant to extract geometric features because the quality of modeling greatly depends on therepresentation of features. Some fitting techniques of natural quadric surfaces with least-squaresmethod are described. And these techniques can be directly used to extract quadric surfaces featuresduring the process of segmentation for point cloud.
基金support provided by The Science and Technology Development Fund,Macao SAR,China(File Nos.0057/2020/AGJ and SKL-IOTSC-2021-2023)Science and Technology Program of Guangdong Province,China(Grant No.2021A0505080009).
文摘Accurate prediction of shield tunneling-induced settlement is a complex problem that requires consideration of many influential parameters.Recent studies reveal that machine learning(ML)algorithms can predict the settlement caused by tunneling.However,well-performing ML models are usually less interpretable.Irrelevant input features decrease the performance and interpretability of an ML model.Nonetheless,feature selection,a critical step in the ML pipeline,is usually ignored in most studies that focused on predicting tunneling-induced settlement.This study applies four techniques,i.e.Pearson correlation method,sequential forward selection(SFS),sequential backward selection(SBS)and Boruta algorithm,to investigate the effect of feature selection on the model’s performance when predicting the tunneling-induced maximum surface settlement(S_(max)).The data set used in this study was compiled from two metro tunnel projects excavated in Hangzhou,China using earth pressure balance(EPB)shields and consists of 14 input features and a single output(i.e.S_(max)).The ML model that is trained on features selected from the Boruta algorithm demonstrates the best performance in both the training and testing phases.The relevant features chosen from the Boruta algorithm further indicate that tunneling-induced settlement is affected by parameters related to tunnel geometry,geological conditions and shield operation.The recently proposed Shapley additive explanations(SHAP)method explores how the input features contribute to the output of a complex ML model.It is observed that the larger settlements are induced during shield tunneling in silty clay.Moreover,the SHAP analysis reveals that the low magnitudes of face pressure at the top of the shield increase the model’s output。
基金This project is supported by Scientific Research Foundation of National Defence of China (No.41319040202).
文摘Aiming to deficiency of the filter and wrapper feature selection methods, anew method based on composite method of filter and wrapper method is proposed. First the methodfilters original features to form a feature subset which can meet classification correctness rate,then applies wrapper feature selection method select optimal feature subset. A successful techniquefor solving optimization problems is given by genetic algorithm (GA). GA is applied to the problemof optimal feature selection. The composite method saves computing time several times of the wrappermethod with holding the classification accuracy in data simulation and experiment on bearing faultfeature selection. So this method possesses excellent optimization property, can save more selectiontime, and has the characteristics of high accuracy and high efficiency.
基金National Nature Science Foundation of China (NO.50275121)
文摘An efficient multi-harmonic method is proposed for studying the effects of mistuning on resonant features of bladed disks with blade-to-blade dry friction damping. This method is able to predict accurately the forced response of bladed disks in frequency domain, which is validated by numerical integration method in time domain. The resonant features of both tuned and mistuned systems are investigated by using this method under various system coupling strengths, viscous dampings, and dry friction darnpings, etc. The results demonstrate that the proposed multi-harmonic method is very efficient for studying the mistuning effects on the resonant response of bladed disks with blade-to-blade dry friction damping, especially considering the combined effects of various system parameters.
基金This project is supported by Provincial Project Foundation of Science and Technology of Guangdong, China(No.2002104040101).
文摘In order to provide a guidance to specify the element size dynamically during adaptive finite element mesh generation, adaptive criteria are firstly defined according to the relationships between the geometrical features and the elements of 3D solid. Various modes based on different datum geometrical elements, such as vertex, curve, surface, and so on, are then designed for generating local refined mesh. With the guidance of the defmed criteria, different modes are automatically selected to apply on the appropriate datum objects to program the element size in the local special areas. As a result, the control information of element size is successfully programmed covering the entire domain based on the geometrical features of 3D solid. A new algorithm based on Delatmay triangulation is then developed for generating 3D adaptive finite element mesh, in which the element size is dynamically specified to catch the geometrical features and suitable tetrahedron facets are selected to locate interior nodes continuously. As a result, adaptive mesh with good-quality elements is generated. Examples show that the proposed method can be successfully applied to adaptive finite element mesh automatic generation based on the geometrical features of 3D solid.
基金This research was funded by the Deputyship for Research&Innovation,Ministry of Education in Saudi Arabia under the Project Number(77/442).
文摘Several millions of people suffer from Parkinson’s disease globally.Parkinson’s affects about 1%of people over 60 and its symptoms increase with age.The voice may be affected and patients experience abnormalities in speech that might not be noticed by listeners,but which could be analyzed using recorded speech signals.With the huge advancements of technology,the medical data has increased dramatically,and therefore,there is a need to apply data mining and machine learning methods to extract new knowledge from this data.Several classification methods were used to analyze medical data sets and diagnostic problems,such as Parkinson’s Disease(PD).In addition,to improve the performance of classification,feature selection methods have been extensively used in many fields.This paper aims to propose a comprehensive approach to enhance the prediction of PD using several machine learning methods with different feature selection methods such as filter-based and wrapper-based.The dataset includes 240 recodes with 46 acoustic features extracted from3 voice recording replications for 80 patients.The experimental results showed improvements when wrapper-based features selection method was used with K-NN classifier with accuracy of 88.33%.The best obtained results were compared with other studies and it was found that this study provides comparable and superior results.
文摘The development of defect prediction plays a significant role in improving software quality. Such predictions are used to identify defective modules before the testing and to minimize the time and cost. The software with defects negatively impacts operational costs and finally affects customer satisfaction. Numerous approaches exist to predict software defects. However, the timely and accurate software bugs are the major challenging issues. To improve the timely and accurate software defect prediction, a novel technique called Nonparametric Statistical feature scaled QuAdratic regressive convolution Deep nEural Network (SQADEN) is introduced. The proposed SQADEN technique mainly includes two major processes namely metric or feature selection and classification. First, the SQADEN uses the nonparametric statistical Torgerson–Gower scaling technique for identifying the relevant software metrics by measuring the similarity using the dice coefficient. The feature selection process is used to minimize the time complexity of software fault prediction. With the selected metrics, software fault perdition with the help of the Quadratic Censored regressive convolution deep neural network-based classification. The deep learning classifier analyzes the training and testing samples using the contingency correlation coefficient. The softstep activation function is used to provide the final fault prediction results. To minimize the error, the Nelder–Mead method is applied to solve non-linear least-squares problems. Finally, accurate classification results with a minimum error are obtained at the output layer. Experimental evaluation is carried out with different quantitative metrics such as accuracy, precision, recall, F-measure, and time complexity. The analyzed results demonstrate the superior performance of our proposed SQADEN technique with maximum accuracy, sensitivity and specificity by 3%, 3%, 2% and 3% and minimum time and space by 13% and 15% when compared with the two state-of-the-art methods.
基金the National Natural Science Foundation of China(No.71401016)the Shaanxi Provincial Natural Science Foundation of China(No.2019JM-495)the Fundamental Research Funds for Central Universities of Chang'an University(Nos.300102228110 and 300102228402)。
文摘Mahalanobis-Taguchi system(MTS)is a kind of data mining and pattern recognition method which can identify the attribute characteristics of multidimensional data by constructing Mahalanobis distance(MD)measurement scale.In this paper,considering the influence of irregular distribution of the sample data and abnormal variation of the normal data on accuracy of MTS,a feature recognition and selection model of the equipment state based on the improved MTS is proposed,and two aspects of the model namely construction of the original Mahalanobis space(MS)and determination of the threshold are studied.Firstly,the original training sample space is statistically controlled by the X-bar-S control chart,and extreme data of the single characteristic attribute is filtered to reduce the impact of extreme condition on the accuracy of the model,so as to construct a more robust MS.Furthermore,the box plot method is used to determine the threshold of the model.And the stability of the model and the tolerance to the extreme condition are improved by leaving sufficient range of the variation for the extreme condition which is identified as in the normal range.Finally,the improved model is compared with the traditional one based on the unimproved MTS by using the data from the literature.The result shows that compared with the traditional model,the accuracy and sensitivity of the improved model for state identification can be greatly enhanced.
基金supported by National Key Research and Development Program of China(No.2017YFB1102800)National Natural Science Foundation of China(Nos.11432011,11620101002,11722219).
文摘This paper presents a multiscale design method for simultaneous topology optimization of both macrostructures and microstructures.Geometric features are extended as design primitives at both macro and micro scales and represented by Level Set Functions(LSFs).Parameters related to the locations,sizes,and orientations of macro and micro features are considered as design variables and optimized simultaneously.In the overlapping areas of different macro features,embedded microstructures are optimally figured out as the solution of the corresponding sub-optimization,problem.In this study,the eXtended Finite Element Method(XFEM)is implemented for structural and sensitivity analyses with respect to design variables.This method has the advantage of using a fixed grid independent of the topology optimization process.The homogenization procedure is applied to calculate the effective properties of considered microstructures in each macro feature.Numerical examples are presented to illustrate the effectiveness of the proposed method.Results depict that the multiscale design cannot obviously improve structural stiffness compared with a solid-material design under the linear elastic condition.
基金Supported by the National Natural Science Foundation of China(60772066)
文摘The feature-selection problem in training AdaBoost classifiers is addressed in this paper. A working feature subset is generated by adopting a novel feature subset selection method based on the partial least square （PLS） regression, and then trained and selected from this feature subset in Boosting. The experiments show that the proposed PLS-based feature-selection method outperforms the current feature ranking method and the random sampling method.
基金supported by the National Natural Science Foundation of China(Grant No.40875012)the National Basic Research Program of China(Grant No.2009CB421502)the Meteorology Open Fund of Huaihe River Basin(HRM200704).
文摘We describe how the Unit-Feature Spatial Classification Method(UFSCM) can be used operationally to classify cloud types in satellite imagery efficiently and conveniently.By using a combination of Interactive Data Language(IDL) and Visual C++(VC) code in combination to extend the technique in three dimensions(3-D),this paper provides an efficient method to implement interactive computer visualization of the 3-D discrimination matrix modification,so as to deal with the bi-spectral limitations of traditional two dimensional(2-D) UFSCM.The case study of cloud-type classification based on FY-2C satellite data (0600 UTC 18 and 0000 UTC 10 September 2007) is conducted by comparison with ground station data, and indicates that 3-D UFSCM makes more use of the pattern recognition information in multi-spectral imagery,resulting in more reasonable results and an improvement over the 2-D method.
基金sponsored by National Key R&D Program of China(2018YFC1504504)Youth Foundation of Yunnan Earthquake Agency(2021K01)Project of Yunnan Earthquake Agency“Chuan bang dai”(CQ3-2021001).
文摘In order to improve the accuracy of building structure identification using remote sensing images,a building structure classification method based on multi-feature fusion of UAV remote sensing image is proposed in this paper.Three identification approaches of remote sensing images are integrated in this method:object-oriented,texture feature,and digital elevation based on DSM and DEM.So RGB threshold classification method is used to classify the identification results.The accuracy of building structure classification based on each feature and the multi-feature fusion are compared and analyzed.The results show that the building structure classification method is feasible and can accurately identify the structures in large-area remote sensing images.
文摘ICSED (Improved Cluster Shade Edge Detection) algorithm and other various methods to accurately and efficiently detect edges on satellite data are presented. Error rate criterion is used to statistically evaluate the performances of these methods in detecting oceanic features for both noise free and noise contaminated AVHRR (Advanced Very High Resolution Radiometer) IR image with Kuroshio. Also, practical experiments in detecting the eddy of Kuroshio with these methods are carried out for comparison. Results show that the ICSED algorithm has more advantages than other methods in detecting mesoscale features of ocean. Finally, the effectiveness of window size of ICSED method to oceanic features detection is quantitatively discussed.
文摘Web data extraction has become a key technology for extracting valuable data from websites.At present,most extraction methods based on rule learning,visual pattern or tree matching have limited performance on complex web pages.Through ana-lyzing various statistical characteristics of HTML el-ements in web documents,this paper proposes,based on statistical features,an unsupervised web data ex-traction method—traversing the HTML DOM parse tree at first,calculating and generating the statistical matrix of the elements,and then locating data records by clustering method and heuristic rules that reveal in-herent links between the visual characteristics of the data recording areas and the statistical characteristics of the HTML nodes—which is both suitable for data records extraction of single-page and multi-pages,and it has strong generality and needs no training.The ex-periments show that the accuracy and efficiency of this method are equally better than the current data extrac-tion method.
基金The project entitled "Urban Active Fault Surveying Project"(143623) funded by the National Development and Roform Commission of China"Active Faults Exploration and Seismic Hazard Assessment in Zibo City"(SD1501) funded by the Department of Science & Technology of Shangdong Province,China
文摘Electrical resistivity tomography (ERT) has been used to experimentally detect shallow buried faults in urban areas in the past a few years, with some progress and experience obtained. According to the results from Olympic Park, Beijing, Shandong Province, Gansu Province and Shanxi Province, we have generalized the method and procedure for inferring the discontinuity of electrical structures (DES) indicating a buried fault in urban areas from resistivity tomograms and its typical electrical features. In general, the layered feature of the electrical structure is first analyzed to preliminarily define whether or not a DES exists in the target area. Resistivity contours in resistivity tomograms are then analyzed from the deep to the shallow. If they extend upward from the deep to the shallow and shape into an integral dislocation, sharp flexure (convergence) or gradient zone, it is inferred that the DES exists, indicating a buried fault. Finally, horizontal tracing is be carried out to define the trend of the DES. The DES can be divided into three types-type AB, ABA and AC. In the present paper, the Zhangdian-Renhe fault system in Zibo city is used as an example to illustrate how to use the method to infer the location and spatial extension of a target fault. Geologic drilling holes are placed based on our research results, and the drilling logs testify that our results are correct. However, the method of this paper is not exclusive and inflexible. It is expected to provide reference and assistance for inferring the shallow buried faults in urban areas from resistivity tomograms in the future.
基金supported by the National Natural Science Foundation of China (Nos.61806107 and 61702135)。
文摘We propose a hierarchical multi-scale attention mechanism-based model in response to the low accuracy and inefficient manual classification of existing oceanic biological image classification methods. Firstly, the hierarchical efficient multi-scale attention(H-EMA) module is designed for lightweight feature extraction, achieving outstanding performance at a relatively low cost. Secondly, an improved EfficientNetV2 block is used to integrate information from different scales better and enhance inter-layer message passing. Furthermore, introducing the convolutional block attention module(CBAM) enhances the model's perception of critical features, optimizing its generalization ability. Lastly, Focal Loss is introduced to adjust the weights of complex samples to address the issue of imbalanced categories in the dataset, further improving the model's performance. The model achieved 96.11% accuracy on the intertidal marine organism dataset of Nanji Islands and 84.78% accuracy on the CIFAR-100 dataset, demonstrating its strong generalization ability to meet the demands of oceanic biological image classification.
基金the National Natural Science Foundation of China [Grant Number 41722108],(Grant Number 91638201)%FWO project:data fusion for image analysis in remote sensing(Grant Number G037115N)
文摘Hyperspectral image provides abundant spectral information for remote discrimination of subtle differences in ground covers.However,the increasing spectral dimensions,as well as the information redundancy,make the analysis and interpretation of hyperspectral images a challenge.Feature extraction is a very important step for hyperspectral image processing.Feature extraction methods aim at reducing the dimension of data,while preserving as much information as possible.Particularly,nonlinear feature extraction methods (e.g.kernel minimum noise fraction (KMNF) transformation) have been reported to benefit many applications of hyperspectral remote sensing,due to their good preservation of high-order structures of the original data.However,conventional KMNF or its extensions have some limitations on noise fraction estimation during the feature extraction,and this leads to poor performances for post-applications.This paper proposes a novel nonlinear feature extraction method for hyperspectral images.Instead of estimating noise fraction by the nearest neighborhood information (within a sliding window),the proposed method explores the use of image segmentation.The approach benefits both noise fraction estimation and information preservation,and enables a significant improvement for classification.Experimental results on two real hyperspectral images demonstrate the efficiency of the proposed method.Compared to conventional KMNF,the improvements of the method on two hyperspectral image classification are 8 and 11％.This nonlinear feature extraction method can be also applied to other disciplines where high-dimensional data analysis is required.
