Component Analysis Network Research Articles

Seismic fault identification in coal mines based on the self-organizing map-gray wolf optimizer-support vector machine algorithm

Accurate fault identification in coal mines is important to improve mine safety and economic benefits. We compare various intelligent algorithms for data preprocessing and optimization and analyze the construction methods of seismic attribute data sets and the performance of intelligent optimization algorithms using fault identification accuracy as the discrimination index to find a better combined model for seismic fault identification. First, the training data set is constructed by mining the fault and nonfault information revealed by the roadway. The distribution characteristics of the seismic attribute data indicate similarities among them, and they are nonlinearly separable. Directly using the attributes to construct the data set, the accuracy of fault identification using the support vector machine (SVM) model is 78.41%. Principal component analysis (PCA) and self-organizing mapping (SOM) neural networks are used to extract effective information and then combined with the SVM classification model, and the accuracy of fault identification is 83.82% and 87.47%, respectively. Compared with the original data and PCA dimensionality reduction data, the accuracy of fault detection is improved by 9.06% and 3.66%, respectively, indicating that SOM can effectively improve the accuracy of fault detection by eliminating similar attributes and reducing the weight of redundant information. Then, through a fixed attribute data set, genetic algorithm (GA), particle swarm optimization (PSO), and gray wolf optimizer (GWO) intelligent optimization algorithms are used to find the optimal kernel function parameter and penalty parameter of the SVM classifier. The accuracy rate of the SOM-GWO-SVM model reaches 91.12%, compared with the SOM-PSO-SVM and SOM-GA-SVM, and the model accuracy is increased by 5.2% and 5.61%, respectively. Compared with PSO and GA, the GWO algorithm has a better global search ability. The identification result of the SOM-GWO-SVM model is closest to the actual fault exposure, especially for the identification of “short” faults and associated faults, which has obvious advantages over the traditional manual interpretation in terms of efficiency and accuracy.

Functional PCA and deep neural networks-based Bayesian inverse uncertainty quantification with transient experimental data

This work focuses on developing an inverse uncertainty quantification (IUQ) process for time-dependent responses, using dimensionality reduction by functional principal component analysis (PCA) and deep neural network (DNN)-based surrogate models. The demonstration is based on the IUQ of TRACE physical model parameters using the FEBA benchmark transient experimental data on peak cladding temperatures during reflooding. The quantity-of-interest (QoI) is time-dependent peak cladding temperature (PCT) profiles. Conventional PCA can hardly represent the data precisely due to the sudden temperature drop at the time of quenching. As a result, a functional alignment method is used to separate the phase and amplitude information in the PCT profiles before conventional PCA is applied for dimensionality reduction. The resulting PC scores are then used to build DNN-based surrogate models to significantly reduce the computational cost in Markov Chain Monte Carlo sampling, while the code/interpolation uncertainty is accounted for using Bayesian neural networks. We compared four IUQ processes with different dimensionality reduction methods and surrogate models. The proposed approach has demonstrated the best performance in reducing the dimensionality of the transient PCT profiles. This approach also produces the posterior distributions of the physical model parameters with which the model predictions have the best agreement with the experimental data.

Predicting stone composition via machine-learning models trained on intra-operative endoscopic digital images

ObjectivesThe aim of this study was to use deep learning (DL) of intraoperative images of urinary stones to predict the composition of urinary stones. In this way, the laser frequency and intensity can be adjusted in real time to reduce operation time and surgical trauma.Materials and methodsA total of 490 patients who underwent holmium laser surgery during the two-year period from March 2021 to March 2023 and had stone analysis results were collected by the stone laboratory. A total of 1658 intraoperative stone images were obtained. The eight stone categories with the highest number of stones were selected by sorting. Single component stones include calcium oxalate monohydrate (W1), calcium oxalate dihydrate (W2), magnesium ammonium phosphate hexahydrate, apatite carbonate (CH) and anhydrous uric acid (U). Mixed stones include W2 + U, W1 + W2 and W1 + CH. All stones have intraoperative videos. More than 20 intraoperative high-resolution images of the stones, including the surface and core of the stones, were available for each patient via FFmpeg command screenshots. The deep convolutional neural network (CNN) ResNet-101 (ResNet, Microsoft) was applied to each image as a multiclass classification model.ResultsThe composition prediction rates for each component were as follows: calcium oxalate monohydrate 99% (n = 142), calcium oxalate dihydrate 100% (n = 29), apatite carbonate 100% (n = 131), anhydrous uric acid 98% (n = 57), W1 + W2 100% (n = 82), W1 + CH 100% ( n = 20) and W2 + U 100% (n = 24). The overall weighted recall of the cellular neural network component analysis for the entire cohort was 99%.ConclusionThis preliminary study suggests that DL is a promising method for identifying urinary stone components from intraoperative endoscopic images. Compared to intraoperative identification of stone components by the human eye, DL can discriminate single and mixed stone components more accurately and quickly. At the same time, based on the training of stone images in vitro, it is closer to the clinical application of stone images in vivo. This technology can be used to identify the composition of stones in real time and to adjust the frequency and energy intensity of the holmium laser in time. The prediction of stone composition can significantly shorten the operation time, improve the efficiency of stone surgery and prevent the risk of postoperative infection.

Joint principal component analysis and BP neural network to measure the development level of modern logistics in China

Joint principal component analysis and BP neural network to measure the development level of modern logistics in China

Redes neurais artificiais para descritores morfológicos em germoplasma de alface biofortificada

ABSTRACT The classification based on morphological descriptors in lettuce is considered a complex activity and proves to be efficient for studying phenotypic characteristics. Therefore, the objective of this study was to analyze the biofortified lettuce germplasm bank at the Universidade Federal de Uberlândia using both conventional morphological descriptors and artificial neural networks. The experiment was conducted in the field. The experimental design employed was a randomized complete block design, consisting of 14 treatments (11 genotypes of mini lettuce, and the cultivars Purpurita, UDI 10.000, and Pira 72) with four replications. Nine morphological descriptors were evaluated. Following the data acquisition, dissimilarity matrix analyses, principal component analysis, dendrogram construction, and artificial neural network (ANN) analyses were performed. The genotypes exhibited phenotypic variability when compared to the parental strains UDI 10.000 and Pira 72. The purple color of the leaves and anthocyanin presence across the entire leaf surface were predominant among the genotypes. Descriptors such as leaf intensity and color, as well as anthocyanin intensity, coloration, and distribution, were the most influential in assessing genetic variability. The Self-Organizing Map (SOM) demonstrated greater sensitivity in discriminating between genotypes compared to the Unweighted Pair Group Method with Arithmetic Mean (UPGMA). While the UPGMA clustering method grouped genotypes into three clusters, the SOM method grouped into five clusters. The use of genetic distance analyses and SOM dendrogram proved to be effective in selecting individuals UFU 215#1, UFU 215#2, UFU 215#6, UFU 215#10, and UFU 215#13, which are clustered with the cultivar UFU Mini Biofort.

Estimates of Soil Organic Carbon in the Ojos de Agua and El Quinillal Forests in the Central Huallaga of Peru

The Peruvian Amazon has experienced large losses of forest cover due to changes in land use, contributing to increases in CO2 in the atmosphere. This study estimated the organic carbon content of forest soil in two forests “Ojos de Agua” and “El Quinillal” in the Central Huallaga of Peru, establishing three types of cover: (i) primary, (ii) intervened, and (iii) deforested. For this purpose, 24 plots of 100 m2 were established and samples were extracted at a depth of 0–20 cm. The effect of the type of forest cover on soil carbon (Organic Carbon‐SOC, Inorganic Carbon‐SIC, Saturated Carbon‐SC, Critical Carbon‐CC, Saturated Carbon Deficit‐SCD, and Organic Carbon‐OC) was analyzed by means of an Analysis of Variance, correlation. Likewise, the relationship between carbon (C) and soil properties was evaluated by principal component analysis and correlation network. The results indicated that the highest SOC averages were found in the primary forests of Ojos de Agua and El Quinillal with 3.54% and 2.51%. The lowest values were found in the deforested forests with 1.34% and 1.46%. The calculation of the saturated C levels of the soil showed an average of 28.63% ± 2.14% and the saturated carbon deficit of 26.63% ± 2.45%, whereas the critical threshold of C showed an average of 2.21% ± 0.18%. The highest SOC content found in the Ojos de Agua primary forest is due to the presence of dominant forest species such as Manilkara bidentata and Brosimum alicastrum. Likewise, the C deficit in the soils of the study area is very alarming because it is very close to saturation levels, especially in deforested forests.

Dysfunction of the triple-network model is associated with cognitive impairment in patients with cerebral small vessel disease

PurposeThis study aimed to demonstrate the correlations between the altered functional connectivity patterns in the triple-network model and cognitive impairment in patients with cerebral small vascular disease (CSVD). MethodsResting-state functional magnetic resonance imaging data were obtained from 22 patients with CSVD and 20 healthy controls. The resting-state data were analyzed using independent component analysis and functional network connectivity (FNC) analysis to explore the functional alterations in the intrinsic triple-network model including the salience network (SN), default mode network (DMN), and central executive network (CEN), and their correlations with the cognitive deficits and clinical observations in the patients with CSVD. ResultsCompared to the healthy controls, the patients with CSVD exhibited increased connectivity patterns in the CEN-DMN and decreased connectivity patterns in the DMN-SN, CEN-SN, intra-SN, and intra-DMN. Significant negative correlations were detected between the intra-DMN connectivity pattern and the Montreal Cognitive Assessment (MoCA) total scores (r = −0.460, p = 0.048) and MoCA abstraction scores (r = −0.565, p = 0.012), and a positive correlation was determined between the intra-SN connectivity pattern and the MoCA abstraction scores (r = 0.491, p = 0.033). ConclusionsOur study findings suggest that the functional alterations in the triple-network model are associated with the cognitive deficits in patients with CSVD and shed light on the importance of the triple-network model in the pathogenesis of CSVD.

Multi-stage Dynamic Latent Variable Model and Lightweight PCANet Based Anomaly Identification of Fused Magnesium Furnaces

The fused magnesium furnace (FMF) is a crucial component in the production of magnesium. The inability to timely detect abnormal operation can result in a semi-molten state, posing a serious risk to production safety. The limitations of current monitoring technologies hinder effective anomaly detection due to their failure to extract intrinsic data features and fully utilize multi-source information. This paper introduces a novel multi-source information fusion approach for anomaly identification of FMFs, leveraging dynamic feature extraction. A two-layer multi-stage dynamic latent variable model is established for dynamic feature extraction. Meanwhile, a lightweight principal component analysis network (PCANet) is established for video deep-learning online identification. The lightweight model can improve the recognition efficiency and reduce the computational cost. Subsequently, through decision-level fusion of multi-source information, the final diagnostic outcomes facilitate the classification and identification of FMF operating conditions. Experiments on real industrial data substantiate the efficacy and advancement of the proposed method. The final accuracy reached 98.43%.

Cancelable face recognition using phase retrieval and complex principal component analysis network

Cancelable face recognition using phase retrieval and complex principal component analysis network

A Study on the Construction of Cultivating High-Quality Farmers in Higher Vocational Colleges and Universities under the OBE-PDCA Approach

Abstract Rural revitalization, talent first, from the high-quality farmer’s cultivation policy as well as vocational education characteristics research, to achieve high-quality farmers’ talent cultivation. This paper analyzes the status quo of the cultivation of high-quality farmers, constructs the cultivation model of high-quality farmers through the OBE-PDCA method, and constructs the cultivation effect evaluation model of high-quality farmers based on the BP-PCA method by combining BP neural network and principal component analysis. On this basis, the cultivation path of management-type high-quality farmers is taken as an example to explore the effect of OBE-PDCA in constructing the cultivation model of high-quality farmers as well as the cultivation effect of high-quality farmers. The results show that the averages of the cultivation of provincial higher vocational colleges and universities in the four dimensions of PDCA are 16.07, 10.51, 13.48, and 22.58, and the F-values of the overall test are 2.4, 7.8, 4.5, and 4.1, with the correlation coefficients of 0.46, 0.45, and 0.59, respectively. The resulting difference in the weight scores of Δ ≥ 10 is 1 item, 5 ≤ Δ ＜ 10 is 6 items, 0≤ Δ ＜ 5 is 1 item, with good cultivation effects. Item, there is a good cultivation effect. The method constructed in this paper can effectively enhance the cultivation effect of high-quality farmers and promote rural revitalization.

HPLC-LED-Induced Fluorescence Analysis of Tear Fluids: An Objective Method for Primary Open Angle Glaucoma Diagnosis

PurposeThe study showcased the application of the lab-assembled HPLC-LED-IF system to analyze proteins in tear fluid samples collected from individuals diagnosed with primary open-angle glaucoma (POAG).MethodsClinical application of the said technique was evaluated by recording chromatograms of tear fluid samples from control and POAG subjects and by analyzing the protein profile using multivariate analysis. The data analysis methods involved are principal component analysis (PCA), Match/No-Match, and artificial neural network (ANN) based binary classification for disease diagnosis.ResultsMahalanobis distance and spectral residual values calculated using a standard calibration set of clinically confirmed POAG samples for the Match/No-Match test gave 86.9% sensitivity and 81.8% specificity. ANN with leaving one out procedure has given 87.1% sensitivity and 81.8% specificity.ConclusionThe results of the study revealed that the utilization of a 278 nm LED excitation in the HPLC system offers good sensitivity for detecting proteins at low concentrations allowing to obtain reliable protein profiles for the diagnosis of POAG.

Development of machine-learning-based models for identifying the sources of nitrate and fluoride in groundwater and predicting their human health risks

AbstractThis research aimed to identify the main sources of groundwater pollution and assess the non-carcinogenic human health risk resulting from nitrate and fluoride contamination. These goals were achieved by employing unsupervised and supervised machine algorithms, including principal component analysis (PCA) and multilayer perceptron artificial neural networks (MLP-ANN). Thirty-seven groundwater samples were analyzed for twelve physical and chemical parameters, including pH, EC, TDS, TH, Cl, F, SO4, NO3, Ca, Mg, Na, and HCO3, and the initial investigation indicated that except for Cl, F, Ca, and Mg, all the parameters are above the guidelines of the World Health Organization (WHO). PCA indicated that mineral dissolution is the main source of F, while high NO3 concentration primarily resulted from agricultural operation due to extensive use of nitrogen and calcium-based fertilizers. Consequently, the non-carcinogenic human health risk (HHR) for children and adults is evaluated based on NO3 and F. The conventional approach for assessing HHR is time-consuming and often associated with errors in calculating hazard quotients (HQ) and hazard indices (HI). In this research, MLP-ANN is suggested to overcome these limitations. In the MLP-ANN modeling, the data were divided into two parts training (80%) and validation (20%), with NO3 and F concentration as inputs and HQ and HI as outputs. The performance of the resulting models was tested using root mean square error (RMSE) and coefficient of determination (R2). The model provided a satisfactory result with a maximum RMSE of 4% and R2 higher than 97% for training and validation. As a result, obtained HIs suggested that 97.3% of the groundwater samples in the study area are suitable for human consumption. The non-carcinogenic HHR is successfully assessed using machine learning algorithms, and the results have led to the conclusion that this approach is highly recommended for effectively managing groundwater resources.

Impact of exposure temperature rise on mass mortality of tidal flat pacific oysters

Oysters are a major commercial and ecological fishery resource. Recently, the oyster industry has experienced mass mortality in summer due to environmental factors. Generally, the survival of oysters in aquatic environments is mainly impacted by environmental stressors such as elevated sea temperatures and reduced salinity; however, the stressors impacting tidal flat oysters that are repeatedly exposed to air remain poorly understood. Hence, we studied the relationship between environmental factors and the survival of tidal flat pacific oysters in Incheon, South Korea, where mass mortality is common. Principal component analysis and Bayesian networks revealed that air temperature (in spring and summer) and sea temperature (in summer) are related to oyster production. In habitats of the tidal flat oysters during the summer, high temperatures of 34.7–35.4°C (maximum 47.6°C) were observed for average durations of 0.8–1.9 hours (maximum 3.6 hours). Furthermore, heat waves occurred for up to 12 consecutive days. Results from the multiple stress test showed that when exposed to 45°C (air temperature) for 4 hours per day, the survival rate of oysters was 42.5% after only 2 days and 0% after 6 days. The findings stemming from the field observations and stress tests suggest that high temperatures during emersion may contribute to mass mortality of oysters in summer, indicating a potential threat to oysters due to climate change. To understand the effects of future thermal stress on oysters more accurately, simultaneous long-term trend analyses and field-based observations are required.

Monitoring of polycyclic aromatic hydrocarbons and toxic metals in blue mussels and oysters for human consumption from Danish waters (2005–2019)

Bivalve molluscs have for many years been used as indicator organisms for environmental pollution with various chemical substances. However, limited information is available on the levels of chemical contaminants in mussels and oysters intended for human consumption. Polycyclic aromatic hydrocarbons (PAHs) and toxic metals, such as cadmium (Cd), mercury (Hg) and lead (Pb) have been analysed in edible bivalve samples and the present study provides the results from samples collected in Danish waters from 2005 to 2019. Principal Component Analysis (PCA) and Pattern Recognition Neural Network (PRNN) analysis were applied to data to identify any differences according to organism, locality, and contaminant levels. Differences were confirmed for Pb concentrations being generally higher in mussels than oysters. Based on toxic metals results, the PCA and PRNN analyses made it possible to categorise samples as a mussel or an oyster, as well as make decisions on the origin of a specific sample. All concentrations in mussels and oysters were below the current European Union maximum levels for the contaminants.

Analysis of factors influencing police investigation work based on principal component analysis

Abstract In this paper, the generation and carrier models of criminal psychological traces are synthesized and applied to detect criminal cases by constructing a criminal psychological trace model. Secondly, the psychological state of the interrogated person is identified using the first and second layers of the principal component analysis network, and the psychological characteristics of the interrogated person are extracted based on the grey time series improved principal component analysis (GPCA) method. Finally, the psychological state of the interrogated person was analyzed by the psychological test of the criminal psychological trace model. The results showed that the AUC of the CIT psychological test effect on criminal experience was 0.638, while the CIT psychological test effect using the criminal psychological trace model was 0.875. This indicates that the research method in this paper can better understand the criminal psychological traces in the police investigation and their effects on the interrogated person and improve the effectiveness and accuracy of the police investigation.

Principal component analysis and deep neural networks in modeling the melt flow index of degradable plastics

Principal component analysis and deep neural networks in modeling the melt flow index of degradable plastics

Data-driven real-time prediction for attitude and position of super-large diameter shield using a hybrid deep learning approach

The presented research introduces a novel hybrid deep learning approach for the dynamic prediction of the attitude and position of super-large diameter shields - a critical consideration for construction safety and tunnel lining quality. This study proposes a hybrid deep learning approach for predicting dynamic attitude and position prediction of super-large diameter shield. The approach consists of principal component analysis (PCA) and temporal convolutional network (TCN). The former is used for employing feature level fusion based on features of the shield data to reduce uncertainty, improve accuracy and the data effect, and 9 sets of required principal component characteristic data are obtained. The latter is adopted to process sequence data in predicting the dynamic attitude and position for the advantages and potential of convolution network. The approach’s effectiveness is exemplified using data from a tunnel construction project in China. The obtained results show remarkable accuracy in predicting the global attitude and position, with an average error ratio of less than 2 mm on four shield outputs in 97.30% of cases. Moreover, the approach displays strong performance in accurately predicting sudden fluctuations in shield attitude and position, with an average prediction accuracy of 89.68%. The proposed hybrid model demonstrates superiority over TCN, long short-term memory (LSTM), and recurrent neural network (RNN) in multiple indexes. Shapley additive exPlanations (SHAP) analysis is also performed to investigate the significance of different data features in the prediction process. This study provides a real-time warning for the shield driver to adjust the attitude and position of super-large diameter shields.

Activation Loop Plasticity and Active Site Coupling in the MAP Kinase, ERK2

Previous studies of the protein kinase, ERK2, using NMR and hydrogen-exchange measurements have shown changes in dynamics accompanying its activation by phosphorylation. However, knowledge about the conformational motions involved is incomplete. Here, we examined ERK2 using long conventional molecular dynamics (MD) simulations starting from crystal structures of phosphorylated (2P) and unphosphorylated (0P) forms. Individual trajectories were run for (5 to 25) μs, totaling 727 μs. The results show unexpected flexibility of the A-loop, with multiple long-lived (>5 μs) conformational states in both 2P- and 0P-ERK2. Differential contact network and principal component analyses reveal coupling between the A-loop fold and active site dynamics, with evidence for conformational selection in the kinase core of 2P-ERK2 but not 0P-ERK2. Simulations of 2P-ERK2 show A-loop states corresponding to restrained dynamics within the N-lobe, including regions around catalytic residues. One A-loop conformer forms lasting interactions with the L16 segment, leading to reduced RMSF and greater compaction in the active site. By contrast, simulations of 0P-ERK2 reveal excursions of A-loop residues away from the C-lobe, leading to greater active site mobility. Thus, the A-loop in ERK2 switches between distinct conformations that reflect coupling with the active site, possibly via the L16 segment. Crystal packing interactions suggest that lattice contacts with the A-loop may restrain its structural variation in X-ray structures of ERK2. The novel conformational states identified by MD expand our understanding of ERK2 regulation, by linking the activated state of the kinase to reduced dynamics and greater compaction surrounding the catalytic site.

Automated atrial fibrillation and ventricular fibrillation recognition using a multi-angle dual-channel fusion network

Atrial fibrillation (AFIB) and ventricular fibrillation (VFIB) are two common cardiovascular diseases that cause numerous deaths worldwide. Medical staff usually adopt long-term ECGs as a tool to diagnose AFIB and VFIB. However, since ECG changes are occasionally subtle and similar, visual observation of ECG changes is challenging. To address this issue, we proposed a multi-angle dual-channel fusion network (MDF-Net) to automatically recognize AFIB and VFIB heartbeats in this work. MDF-Net can be seen as the fusion of a task-related component analysis (TRCA)-principal component analysis (PCA) network (TRPC-Net), a canonical correlation analysis (CCA)-PCA network (CPC-Net), and the linear support vector machine-weighted softmax with average (LS-WSA) method. TRPC-Net and CPC-Net are employed to extract deep task-related and correlation features, respectively, from two-lead ECGs, by which multi-angle feature-level information fusion is realized. Since the convolution kernels of the above methods can be directly extracted through TRCA, CCA and PCA technologies, their training time is faster than that of convolutional neural networks. Finally, LS-WSA is employed to fuse the above features at the decision level, by which the classification results are obtained. In distinguishing AFIB and VFIB heartbeats, the proposed method achieved accuracies of 99.39 % and 97.17 % in intra- and inter-patient experiments, respectively. In addition, this method performed well on noisy data and extremely imbalanced data, in which abnormal heatbeats are much less than normal heartbeats. Our proposed method has the potential to be used as a diagnostic tool in the clinic.

Assessing thermoelectric performance of quasi 0D carbon and polyaniline nanocomposites using machine learning

Thermoelectric materials have been widely recognized as a simple approach to harness green energy by converting thermal gradients into electrical energy. However, the intricate interplay between electrical conductivity, Seebeck coefficient, and thermal conductivity in thermoelectric materials presents a challenge to improving their efficiency. Traditional experimental methods and calculation methods have troublesome steps and long cycles for predicting new thermoelectric materials. In this work we present materials informatics-based approach, where statistical and machine learning models like correlation matrix, multiple linear regression, principal component analysis and artificial neural network were employed to find the relationship between features and thermoelectric performance. Furthermore, artificial neural network was used to analyze the roles of several compositional and microstructural features along with temperature on electrical conductivity, thermal conductivity, Seebeck coefficient and thermoelectric figure of merit (ZT) for PANI and quasi 0D carbon-based composites.