Shorter Training Time Research Articles

Spatial Network Decomposition for Fast and Scalable AC-OPF Learning

This paper proposes a novel machine-learning approach for predicting AC-OPF solutions that features a fast and scalable training. It is motivated by the significant training time needed by existing machine-learning approaches for predicting AC-OPF. The proposed approach is a 2-stage methodology that exploits a spatial decomposition of the power network that is viewed as a set of regions. The first stage learns to predict the flows and voltages on the buses and lines coupling the regions, and the second stage trains, in parallel, the machine-learning models for each region. The predictions can then seed a power flow to eliminate the physical constraint violations, resulting in minor violations only for the operational bound constraints. Experimental results on the French transmission system (up to 6,700 buses) and large test cases from the <monospace xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">pglib</monospace> library (up to 9,000 buses) demonstrate the potential of the approach. Within a short training time, the approach predicts AC-OPF solutions with very high fidelity, producing significant improvements over the state-of-the-art. The proposed approach thus opens the possibility of training machine-learning models quickly to respond to changes in operating conditions.

Performance of Multiple Pretrained BERT Models to Automate and Accelerate Data Annotation for Large Datasets.

To develop and evaluate domain-specific and pretrained bidirectional encoder representations from transformers (BERT) models in a transfer learning task on varying training dataset sizes to annotate a larger overall dataset. The authors retrospectively reviewed 69 095 anonymized adult chest radiograph reports (reports dated April 2020-March 2021). From the overall cohort, 1004 reports were randomly selected and labeled for the presence or absence of each of the following devices: endotracheal tube (ETT), enterogastric tube (NGT, or Dobhoff tube), central venous catheter (CVC), and Swan-Ganz catheter (SGC). Pretrained transformer models (BERT, PubMedBERT, DistilBERT, RoBERTa, and DeBERTa) were trained, validated, and tested on 60%, 20%, and 20%, respectively, of these reports through fivefold cross-validation. Additional training involved varying dataset sizes with 5%, 10%, 15%, 20%, and 40% of the 1004 reports. The best-performing epochs were used to assess area under the receiver operating characteristic curve (AUC) and determine run time on the overall dataset. The highest average AUCs from fivefold cross-validation were 0.996 for ETT (RoBERTa), 0.994 for NGT (RoBERTa), 0.991 for CVC (PubMedBERT), and 0.98 for SGC (PubMedBERT). DeBERTa demonstrated the highest AUC for each support device trained on 5% of the training set. PubMedBERT showed a higher AUC with a decreasing training set size compared with BERT. Training and validation time was shortest for DistilBERT at 3 minutes 39 seconds on the annotated cohort. Pretrained and domain-specific transformer models required small training datasets and short training times to create a highly accurate final model that expedites autonomous annotation of large datasets.Keywords: Informatics, Named Entity Recognition, Transfer Learning Supplemental material is available for this article. ©RSNA, 2022See also the commentary by Zech in this issue.

Feature selection approach for evolving reactive scheduling policies for dynamic job shop scheduling problem using gene expression programming

Dispatching rules are one of the most widely applied methods for solving Dynamic Job Shop Scheduling problems (DJSSP) in real-world manufacturing systems. Hence, the automated design of effective rules has been an important subject in the scheduling literature for the past several years. High computational requirements and difficulty in interpreting generated rules are limitations of literature methods. Also, feature selection approaches in the field of automated design of scheduling policies have been developed for the tree-based GP approach only. Therefore, the aim of this study is to propose a feature selection approach for the Gene Expression Programming (GEP) algorithm to evolve high-quality rules in simple structures with an affordable computational budget. This integration speeds up the search process by restricting the GP search space using the linear representation of the GEP algorithm and creates concise rules with only meaningful features using the feature selection approach. The proposed algorithm is compared with five algorithms and 30 rules from the literature under different processing conditions. Three performance measures are considered including total weighted tardiness, mean tardiness, and mean flow time. The results show that the proposed algorithm can generate smaller rules with high interpretability in a much shorter training time.

Using Machine Learning Algorithms For Classifying Transmission Line Faults

The faults in transmission lines should be identified for attaining high quality energy in electrical power systems. Savings can be made in both time and energy if the transmission line faults are classified accurately. The present study examined phase-ground, phase-phase-ground, phase-phase, phase-phase-phase and no fault cases. Support Vector Machine (SVM), K-Nearest Neighbours Algorithm (KNN), Decision Tree (DT), Ensemble, Linear discriminant analysis (LDA) classifiers were used for classifying the transmission line faults. These algorithms were compared with regard to parameters such as accuracy, error rate, prediction speed and training time. The accuracy and minimum error of SVM and KNN classifiers were 99.7 % and 0.0011 respectively. DT classifier is faster than the other classifiers with a predicted speed of 29000 obs/sec. Whereas LDA had the shortest training time of 0.76992 sec. The results have indicated that SVM, KNN classifiers have similar performances. In addition, the classifiers SVM, KNN acquired minimum error with the highest accuracy compared with the other classifiers. While DT has the highest estimation speed, LDA has the shortest training time.

A Prognostic and Health Management Framework for Aero-Engines Based on a Dynamic Probability Model and LSTM Network

In this study, a prognostics and health management (PHM) framework is proposed for aero-engines, which combines a dynamic probability (DP) model and a long short-term memory neural network (LSTM). A DP model based on Gaussian mixture model-adaptive density peaks clustering algorithm, which has the advantages of an extremely short training time and high enough precision, is employed for modelling engine fault development from the beginning of engine service, and principal component analysis is introduced to convert complex high-dimensional raw data into low-dimensional data. The model can be updated from time to time according to the accumulation of engine data to capture the occurrence and evolution process of engine faults. In order to address the problems with the commonly used data driven methods, the DP + LSTM model is employed to estimate the remaining useful life (RUL) of the engine. Finally, the proposed PHM framework is validated experimentally using NASA’s commercial modular aero-propulsion system simulation dataset, and the results indicate that the DP model has higher stability than the classical artificial neural network method in fault diagnosis, whereas the DP + LSTM model has higher accuracy in RUL estimation than other classical deep learning methods.

Construction of a Semantic Segmentation Network for the Overhead Catenary System Point Cloud Based on Multi-Scale Feature Fusion

Accurate semantic segmentation results of the overhead catenary system (OCS) are significant for OCS component extraction and geometric parameter detection. Actually, the scenes of OCS are complex, and the density of point cloud data obtained through Light Detection and Ranging (LiDAR) scanning is uneven due to the character difference of OCS components. However, due to the inconsistent component points, it is challenging to complete better semantic segmentation of the OCS point cloud with the existing deep learning methods. Therefore, this paper proposes a point cloud multi-scale feature fusion refinement structure neural network (PMFR-Net) for semantic segmentation of the OCS point cloud. The PMFR-Net includes a prediction module and a refinement module. The innovations of the prediction module include the double efficient channel attention module (DECA) and the serial hybrid domain attention (SHDA) structure. The point cloud refinement module (PCRM) is used as the refinement module of the network. DECA focuses on detail features; SHDA strengthens the connection of contextual semantic information; PCRM further refines the segmentation results of the prediction module. In addition, this paper created and released a new dataset of the OCS point cloud. Based on this dataset, the overall accuracy (OA), F1-score, and mean intersection over union (MIoU) of PMFR-Net reached 95.77%, 93.24%, and 87.62%, respectively. Compared with four state-of-the-art (SOTA) point cloud deep learning methods, the comparative experimental results showed that PMFR-Net achieved the highest accuracy and the shortest training time. At the same time, PMFR-Net segmentation performance on S3DIS public dataset is better than the other four SOTA segmentation methods. In addition, the effectiveness of DECA, SHDA structure, and PCRM was verified in the ablation experiment. The experimental results show that this network could be applied to practical applications.

Modulation format recognition with transfer learning assisted convolutional neural network using multiple Stokes sectional plane image in multi-core fibers.

A modulation format recognition (MFR) scheme based on multi-core fiber (MCF) is proposed for the next generation of elastic optical networks (EONs). In this scheme, multiple Stokes sectional planes images are used as signal features which are typed into a transfer learning (TL) assisted convolutional neural network (CNN) to realize MFR. Compared with the traditional Jones matrix, the Stokes space mapping method is insensitive to polarization mixing, carrier frequency skew and phase offset, therefore, it has better feature representation ability. TL is introduced to transfer the model used in standard single-mode fiber (SSMF) to MCF transmission, reducing the required training data and complexity. In addition, multiple Stokes sectional planes images are input simultaneously, which improves the accuracy of the neural network. Experimental verifications were performed for a polarization division multiplexing (PDM)-EONs system at a symbol rate of 12.5GBaud by 5 km MCF. Nine modulation formats, including three standard modulation formats (BPSK, QPSK, 8PSK), three uniformly shaped (US) modulation formats (US-8QAM, US-16QAM, US-32QAM) and three probabilistically shaped (PS) modulation formats (PS-8QAM, PS-16QAM, PS-32QAM), were recognized by our scheme. The experimental results show that the scheme achieves high recognition accuracy even at low optical signal-to-noise ratio (OSNR). Moreover, the required number of training samples is less 40% compared to the traditional CNN. The proposed scheme has a high tolerance to the crosstalk damage of MCF itself and can realize the short training time of large-capacity space division multiplexing (SDM)-EONs. Our findings have the potential to be used in the next generation of a SDM fiber transmission system.

Performance evaluation of deep transfer learning on multi-class identification of common weed species in cotton production systems

Precision weed management offers a promising solution for sustainable cropping systems through the use of chemical-reduced/non-chemical robotic weeding techniques, which apply suitable control tactics to individual weeds or small clusters. Therefore, accurate identification of weed species plays a crucial role in such systems to enable precise, individualized weed treatment. Despite recent progress, the development of a robust weed identification and localization system in the presence of unstructured field environments remains a serious challenge, requiring supervised modeling using large volumes of annotated data. This paper makes a first comprehensive evaluation of deep transfer learning (DTL) for identifying common weed species specific to cotton (Gossypium hirsutum L.) production systems in southern United States (U.S.). A new dataset for weed identification was created, consisting of 5187 color images of 15 weed classes collected under natural light conditions and at varied weed growth stages, in cotton fields (primarily in Mississippi and North Carolina) during the 2020 and 2021 growth seasons. We evaluated 35 state-of-the-art deep learning models through transfer learning with repeated holdout validations and established an extensive benchmark for the considered weed identification task. DTL achieved high classification accuracy of F1 scores exceeding 95%, requiring reasonably short training time (less than 2.5 h) across models. ResNeXt101 achieved the best overall F1-score of 98.93 ± 0.34%, whereas 10 out of the 35 models achieved F1 scores near or above 98.0%. However, the performance on minority weed classes with few training samples was less satisfactory for models trained with a conventional, unweighted cross entropy loss function. To address this issue, a weighted cross entropy loss function was adopted, which achieved substantially improved accuracies for minority weed classes (e.g., the F1-scores for Xception and MnasNet on the Spurred Anoda weed increased from 48% to 90% and 50% to 82%, respectively). Furthermore, a deep learning-based cosine similarity metric was employed to analyze the similarity among weed classes, assisting in the interpretation of classifications. Both the codes (https://github.com/Derekabc/CottonWeeds) for model benchmarking and the weed dataset (https://www.kaggle.com/yuzhenlu/cottonweedid15) of this study are made publicly available, which expect to be a valuable resource for future research on weed identification and beyond.

Patch-based artifact reduction for three-dimensional volume projection data of sparse-view micro-computed tomography.

Micro-computed tomography (micro-CT) enables the non-destructive acquisition of three-dimensional (3D) morphological structures at the micrometer scale. Although it is expected to be used in pathology and histology to analyze the 3D microstructure of tissues, micro-CT imaging of tissue specimens requires a long scan time. A high-speed imaging method, sparse-view CT, can reduce the total scan time and radiation dose; however, it causes severe streak artifacts on tomographic images reconstructed with analytical algorithms due to insufficient sampling. In this paper, we propose an artifact reduction method for 3D volume projection data from sparse-view micro-CT. Specifically, we developed a patch-based lightweight fully convolutional network to estimate full-view 3D volume projection data from sparse-view 3D volume projection data. We evaluated the effectiveness of the proposed method using physically acquired datasets. The qualitative and quantitative results showed that the proposed method achieved high estimation accuracy and suppressed streak artifacts in the reconstructed images. In addition, we confirmed that the proposed method requires both short training and prediction times. Our study demonstrates that the proposed method has great potential for artifact reduction for 3D volume projection data under sparse-view conditions.

Damage Detection in Wind Turbine Blades Based on an Improved Broad Learning System Model

The research on damage detection in wind turbine blades plays an important role in reducing the risk of shut down in wind turbines. Rapid and accurate damage identification by using efficient detection models is the focus of the current research on damage detection in wind turbine blades. To solve the problems of the complex structure of the model and high time consumption in deep learning models, an improved broad learning system (BLS) model using the algorithm of chunking based on non-local means (NLMs) was proposed, which was called the CBNLM-BLS. The chunked, in-parallel accelerated integral image approach was used to optimize the NLM to speed up the BLS. Experiment results showed that the proposed model achieved a classification accuracy of 99.716%, taking 28.662 s to detect damage in the wind turbine blades. Compared with deep neural network models, such as ResNet, AlexNet and VGG-19, the proposed CBNLM-BLS had higher classification accuracy, shorter training time and less complex model construction and parameters. Compared with traditional BLSs, the CBNLM-BLS had less time complexity. It is of great significance to identify damage in wind turbine blades more efficiently.

Trainable segmentation for transmission electron microscope images of inorganic nanoparticles.

We present a trainable segmentation method implemented within the python package ParticleSpy. The method takes user labelled pixels, which are used to train a classifier and segment images of inorganic nanoparticles from transmission electron microscope images. This implementation is based on the trainable Waikato Environment for Knowledge Analysis (WEKA) segmentation, but is written in python, allowing a large degree of flexibility and meaning it can be easily expanded using other python packages. We find that trainable segmentation offers better accuracy than global or local thresholding methods and requires as few as 100 user-labelled pixels to produce an accurate segmentation. Trainable segmentation presents a balance of accuracy and training time between global/local thresholding and neural networks, when used on transmission electron microscope images of nanoparticles. We also quantitatively investigate the effectiveness of the components of trainable segmentation, its filter kernels and classifiers, in order to demonstrate the use cases for the different filter kernels in ParticleSpy and the most accurate classifiers for different data types. A set of filter kernels is identified that are effective in distinguishing particles from background but that retain dissimilar features. In terms of classifiers, we find that different classifiers perform optimally for different image contrast; specifically, a random forest classifier performs best for high-contrast ADF images, but that QDA and Gaussian Naïve Bayes classifiers perform better for low-contrast TEM images.

Application of machine learning in the identification of fluvial-lacustrine lithofacies from well logs: A case study from Sichuan Basin, China

The lithofacies identification is critical for forecasting sweet spots of hydrocarbon explorations. Well logs are widely used in lithofacies identifications because they are petrophysical measurements of subsurface stratigraphy which reflect lithological successions and depositional processes. The traditional lithofacies identification from well logs is a manual work that is time-consuming and bias-prone. An automated and bias-free method is in demand. To this end, we created a lithofacies dataset of eleven wells with well log records and lithofacies descriptions that were interpreted manually based on facies analysis of drilling cutting descriptions and well logs. Then we developed machine learning models that were trained using the lithofacies dataset of the fluvial-lacustrine Upper Triassic Xujiahe and Lower Jurassic Ziliujing formations in Yuanba Area, northern Sichuan Basin of southwestern China. By employing extreme gradient boosting and resampling algorithms, this machine learning model is efficient and outperforms support vector machine and multiple-layer perceptron, as indicated by its highest accuracy and F1-score of 0.90, the highest AUC of 0.94, as well as the shortest training time. Moreover, the result suggests that resampling is necessary for lithofacies identification with the imbalanced dataset. A combined method of oversampling and undersampling is better than a single resampling method. This study presents a successful application of machine learning in fluvial-lacustrine lithofacies identification from well logs and suggests the great potentiality of machine learning in subsurface hydrocarbon explorations.

Preclinical Validation of a Semi-Autonomous Robot for Transperineal Prostate Biopsy

Prostate biopsy is a manual procedure carried out mostly under ultrasound (US) guidance to confirm the presence of cancer. The standard biopsy is random and includes at least 12 insertions; targeted biopsy makes use of dedicated hardware and software, but is still performed manually. We present here the pre-clinical validation of PROST, a robot primarily designed to automate targeted transperineal biopsy. The overall validation of the system was performed on cadavers, while some features, such as image segmentation, were tested on human tissue. PROST is designed to minimize human error by introducing some autonomy in the execution of key steps of the procedure, i.e., target selection, image fusion and needle positioning. The protocol was approved by the ethics committee; 10 cadavers were included in the study. We envision that PROST has the potential to increase the detection of clinically significant prostate cancer, to simplify the procedure, to reduce human errors and to shorten training time. The use of a robot for the biopsy of the prostate will create the possibility to include also a treatment, such as focal ablation, to be delivered through the same system.

Fault identification of fan blade based on improved ResNet-18

With the continuous development of the wind industry, wind turbine blade failure causes wind turbine shutdown, which damages the revenue of wind power. In order to more effectively detect the damage on the surface of wind turbine blade. By comparing ResNet-18 /50, VGG-19 and GoogLeNet models, ResNet-18 has the advantages of short training time and high accuracy. Therefore, ResNet-18 convolutional neural network is selected as the basic network, and a deep learning training and detection model is built based on MATLAB. In order to further optimize the model, multiple convolution kernel is selected for the main branch of residual structure of ResNet-18 to extract features, and add 3×3 pool layer, it is used to improve the learning ability of the model. In addition, Aadm, Sgdm and Rmsprop optimization algorithms are added to further improve the robustness of the model. Experiments show that the classification and recognition accuracy of the improved ResNet-18 model for fan blades is 94.17%, which is 2.5% higher than that of the original model. The improved ResNet-18 model has good recognition effect,it improves a faster and more accurate solution for fan blade surface damage identification in complex weather environment.

Wood species recognition from wood images with an improved CNN1

With the development of convolutional neural networks, aiming at the problem of low efficiency and low accuracy in the process of wood species recognition, a recognition method using an improved convolutional neural network is proposed in this article. First, a large-scale wood dataset was constructed based on the WOOD-AUTH dataset and the data collected. Then, a new model named W_IMCNN was constructed based on Inception and mobilenetV3 networks for wood species identification. Experimental results showed that compared with other models, the proposed model had better recognition performance, such as shorter training time and higher recognition accuracy. In the data set constructed by us, the accuracy of the test set reaches 96.4%. We used WOOD-AUTH dataset to evaluate the model, and the recognition accuracy reached 98.8%. Compared with state-of-the-art methods, the effectiveness of the W_IMCNN were confirmed.

Vortex-induced vibration fatigue damage prediction method for flexible cylinders based on RBF neural network

Vortex-induced vibration (VIV) is an important factor resulting in fatigue failure of cylindrical structures in offshore engineering applications (such as risers and cables). The traditional empirical model for VIV prediction often yields unsatisfactory results due to its dependence on the selection of empirical coefficients. Thus, establishing a general approach to VIV damage prediction based on artificial neural networks and engineering experience is of great significance to engineering guidance. In this paper, a VIV damage prediction model of a flexible cylinder was established using a radial basis function neural network (RBFNN) and model test data. Compared with back propagation neural network (BPNN), genetic algorithm-back propagation neural network (GA-BPNN), and generalized regression neural network (GRNN), the RBFNN exhibited excellent performance with high accuracy and the shortest training time in predicting the VIV damage of flexible cylinders. In addition, the VIV fatigue of flexible cylinders fitted with passive control devices (control rods and helical strakes) was predicted. Using three and four control rods, the VIV reduction effect was observed to be better when the attack angle θ was in the range of 30°–45° and 15°–30°, respectively. Further, regarding helical strakes, the damage values of the cylinder were significantly reduced; however, the VIV suppression effect was significantly affected by the yaw angles. The VIV damage prediction model of this paper is expected to serve as important engineering guidance for the preliminary design, installation arrangement, operation, and maintenance of flexible cylindrical structures in offshore engineering applications.

Investigating the Effect of Flickering Frequency Pair and Mother Wavelet Selection in Steady-State Visually-Evoked Potentials on Two-Command Brain-Computer Interfaces

IntroductionSteady-state visually evoked potentials (SSVEPs) have become popular in brain-computer interface (BCI) applications in addition to many other applications on clinical neuroscience (neurodegenerative disorders, schizophrenia, epilepsy, etc.), cognitive (visual attention, working memory, brain rhythms, etc.), and use of engineering researches. Among available methods to measure brain activities, SSVEPs have advantages like higher information transfer rate, simplicity in structure, and short training time. SSVEP-based BCIs use flickering stimuli at different frequencies to discriminate distinct commands in real life. Some features are extracted from the SSVEP signals before these commands are classified. The wavelet transform (WT) has attracted researchers among feature extraction methods since it utilizes the non-stationary signals well. In the WT, a sample function (named mother wavelet) represents the SSVEP signal in both time and frequency domains. Unfortunately, there is no universal mother wavelet function that fits all the signals. Therefore, choosing an appropriate mother wavelet function may be a challenge in WT-related studies. Although there are such studies in three- and seven-command SSVEP-based studies, there is no study for two-command systems in our knowledge. Materials and MethodsIn this study, two user commands flickered at the combinations of seven different frequencies were tested to determine which frequency pairs give the highest performance. For this purpose, three well-known wavelet features (energy, entropy, and variance) were calculated for each of derived EEG frequency bands from the discrete WT coefficients of SSVEP signals. The WT was repeated for six different mother wavelet functions (Haar, Db4, Sym4, Coif1, Bior3.5, and Rbior2.8). Then, four feature sets (every three features, and all together) were applied to seven commonly-used machine learning algorithms (Decision Tree, Discriminant Analysis, Logistic Regression, Naive Bayes, Support Vector Machines, Nearest Neighbors, and Ensemble Classifiers). Results and DiscussionWe achieved 100% accuracies among these 3,528 runs (7 classifiers x 4 feature sets x 6 mother wavelets x 21 flickering frequency pairs) using the mother wavelet function of Haar and the Ensemble Learner classifier. The highest classifier performances are 100% when two commands have the flickering frequency pairs of (6.0 and 10 Hz), (6.5 and 8.2 Hz), or (6.5 and 10.0 Hz). ConclusionWe obtained three main outcomes from this study. First, the most representative mother wavelet function was Haar, while the worst one was Symlet 4. Second, the Ensemble Learner classifier gave the maximum classifier performance in a two-command SSVEP-based BCI system. Besides, two user commands from SSVEP should be one of the frequency pairs of (6.0 and 10.0 Hz), (6.5 and 8.2 Hz), and (6.5 and 10.0 Hz) to achieve the maximum accuracy.

A review of researches on decoding algorithms of steady-state visual evoked potentials

Brain-computer interface (BCI) systems based on steady-state visual evoked potential (SSVEP) have become one of the major paradigms in BCI research due to their high signal-to-noise ratio and short training time required by users. Fast and accurate decoding of SSVEP features is a crucial step in SSVEP-BCI research. However, the current researches lack a systematic overview of SSVEP decoding algorithms and analyses of the connections and differences between them, so it is difficult for researchers to choose the optimum algorithm under different situations. To address this problem, this paper focuses on the progress of SSVEP decoding algorithms in recent years and divides them into two categories-trained and non-trained-based on whether training data are needed. This paper also explains the fundamental theories and application scopes of decoding algorithms such as canonical correlation analysis (CCA), task-related component analysis (TRCA) and the extended algorithms, concludes the commonly used strategies for processing decoding algorithms, and discusses the challenges and opportunities in this field in the end.

Using Improved BPNN to Predict the Time of Social Poverty Reduction under the Background of Internet

The world has been overturned by the power of the Internet, and all walks of life are constantly changing with the trend of the times. Therefore, with the help of the effective technology and means of the Internet, people can start from a more complex perspective. Analyze the social environment and economic situation, and focus on poverty. By actively monitoring the number of poor people in society, we can grasp the actual situation of poverty. Starting from the depths of the problem, aiming at the characteristics of poor groups, this paper excavates the scale data of poor people. Finally, the improved BP neural network model is used to predict the time when the society gets rid of poverty and reduces poverty. The results show the following: (1) After P value judgment and T value test, the experiment determined 12 indicators and proportion of poverty. The most prominent indicators of contribution rate are T2, T4, and T5. The most significant changes in T test results were T8 and T12. (2) The reformed BP network has less iteration times, shorter training time, and higher prediction accuracy. The network converges 30 times, and the fitting accuracy reaches the best, and the ideal state is 0.000936578. Training the poverty alleviation data from 2016 to 2020 with the model, the final poverty alleviation work has achieved remarkable results. (3) According to the hidden layer nodes of neural network structure, the final number of hidden layer nodes is determined to be 11. (4) Comparing the absolute error and APE error of the prediction results. The predicted value is very close to the actual value, which can correctly reflect the mapping relationship of poverty reduction and the result is reliable. (5) Compare the predicted values of BPNN model before and after modification; it is found that the new BPNN model is more accurate in prediction. The predicted population is close to the actual population, and the error level curve approaches 0%. After empirical test, the improved BPNN model is accurate and effective. Through the poverty reduction prediction of the model, people can observe the actual situation more intuitively. By evaluating the effectiveness of social poverty alleviation, we can optimize the next poverty alleviation strategy and adjust the strategic policy.

COV-ELM classifier: An extreme learning machine based identification of COVID-19 using chest X-ray images

Coronaviruses constitute a family of viruses that gives rise to respiratory diseases. COVID-19 is an infectious disease caused by a newly discovered coronavirus also termed Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2). As COVID-19 is highly contagious, early diagnosis of COVID-19 is crucial for an effective treatment strategy. However, the reverse transcription-polymerase chain reaction (RT-PCR) test which is considered to be a gold standard in the diagnosis of COVID-19 suffers from a high false-negative rate. Therefore, the research community is exploring alternative diagnostic mechanisms. Chest X-ray (CXR) image analysis has emerged as a feasible and effective diagnostic technique towards this objective. In this work, we propose the COVID-19 classification problem as a three-class classification problem to distinguish between COVID-19, normal, and pneumonia classes. We propose a three-stage framework, named COV-ELM based on extreme learning machine (ELM). Our dataset comprises CXR images in a frontal view, namely Posteroanterior (PA) and Erect anteroposterior (AP). Stage one deals with preprocessing and transformation while stage two deals with feature extraction. These extracted features are passed as an input to the ELM at the third stage, resulting in the identification of COVID-19. The choice of ELM in this work has been motivated by its faster convergence, better generalization capability, and shorter training time in comparison to the conventional gradient-based learning algorithms. As bigger and diverse datasets become available, ELM can be quickly retrained as compared to its gradient-based competitor models. We use 10-fold cross-validation to evaluate the results of COV-ELM. The proposed model achieved a macro average F1-score of 0.95 and the overall sensitivity of 0.94 ± 0.02 at a 95% confidence interval. When compared to state-of-the-art machine learning algorithms, the COV-ELM is found to outperform its competitors in this three-class classification scenario. Further, LIME has been integrated with the proposed COV-ELM model to generate annotated CXR images. The annotations are based on the superpixels that have contributed to distinguish between the different classes. It was observed that the superpixels correspond to the regions of the human lungs that are clinically observed in COVID-19 and Pneumonia cases.