Hidden Layer Nodes Research Articles

Artificial Intelligence Algorithm-Based Ultrasound Image Segmentation Technology in the Diagnosis of Breast Cancer Axillary Lymph Node Metastasis.

This paper aimed to investigate the application of ultrasound image segmentation technology based on the back propagation neural network (BPNN) artificial intelligence algorithm in the diagnosis of breast cancer axillary lymph node metastasis, thereby providing a theoretical basis for clinical diagnosis. In this study, 90 breast cancer patients with axillary lymph node metastasis were selected as the research objects and rolled randomly into an experimental group and a control group. Besides, all of them were examined by ultrasound. The BPNN algorithm for the ultrasound image segmentation diagnosis method was applied to the patiens from the experimental group, while the control group was given routine ultrasound diagnosis. Thus, the value of this algorithm in ultrasonic diagnosis was compared and explored. The results showed that when the number of hidden layer nodes based on the BPNN artificial intelligence algorithm was 2, 3, 4, 5, 6, 7, and 8, the corresponding segmentation accuracy was 97.3%, 96.5%, 94.8%, 94.8%, and 94.1% in turn. Among them, the segmentation accuracy was the highest when the number of hidden layer nodes was 2. The correlation of independent variable bubble plot analysis showed that the presence or absence of capsules, the presence of crab feet or burrs in breast cancer lesions was critical influencing factors for the occurrence of axillary lymph node metastasis, and the standardized importance was 99.7% and 70.8%, respectively. Besides, the area under the two-dimensional receiver operating characteristic (ROC) curve of the BPNN artificial intelligence algorithm model classification was always greater than the area under the curve of manual segmentation, and the segmentation accuracy was 90.31%, 94.88%, 95.48%, 95.44%, and 97.65% in sequence. In addition, the segmentation specificity of different running times was higher than that of manual segmentation. In conclusion, the BPNN artificial intelligence algorithm had high accuracy, sensitivity, and specificity for ultrasound image segmentation, with a better segmentation effect. Therefore, it had a better diagnostic effect for breast cancer axillary lymph node metastasis.

HIGH AVAILABILITY IN SERVER CLUSTERS BY USING BACKPROPAGATION NEURAL NETWORK METHOD

Server is a host device applications to serve every request in finding information needs. The server must fully support the services used for the organization's digital needs 24 hours in a day, 7 days in a week, and 365 days in a year. The concept of High Availability is needed to maintain the quality of server services. The algorithm used to build HA can use both classical and modern algorithms. The algorithm used in this research is using backpropagation neural network. In this study, the parameter values to obtain optimal accuracy are learning rate 0.1, training data 80 and test data 20, the number of nodes in hidden layer 4, minimum error 0.0001, and the number of iterations 2500.The best accuracy value using these parameters is 93.79% .

Improving learning in neural networks through weight initializations

Feedforward neural networks are known to exhibit universal approximation property, i.e., these networks can effectively approximate any function arbitrary well, provided there is one hidden layer with sufficient number of hidden nodes, and some non-linearity at the hidden layer nodes. The property of universal approximation in these networks contributed to wide range of application, remarkably in the area of predictive modelling. However, these networks have a major drawback of slow convergence. In this work, we proposed a new weight initialization algorithm to improve the convergence speed in feedforward networks using statistical analysis of system, based on the set of predefined assumptions. The proposed algorithm is compared to other established weight initialization algorithm for predictive models, and is expected to perform better than other algorithms.

Prediction of Compressive Strengths of Concrete with Partial Fine Aggregate of Plastic Using Artificial Neural Network and Revisions

In recent past years, plastic waste has been a environmental menace. Utilization of plastic waste as fine aggregate substitution could reduce the demand and negative impacts of sand mining while addressing waste plastic challenges.
 This study aims at evaluating compressive strengths prediction models for concrete with plastic—mainly recycled plastic—as partial replacement or addition of fine aggregates, by use of artificial neural networks (ANNs), developed in OCTAVE 5.2.0 and datasets from reviews. 44 datasets from 8 different sources were used, that included four input variables namely:- water: binder ratio; control compressive strength (MPa); % plastic replacement or additive by weight and plastic type; and the output variable was the compressive strength of concrete with partial plastic aggregates.
 Various models were run and the selected model, with 14 nodes in hidden layer and 320,000 iterations, indicated overall root mean square error (RMSE) , absolute factor of variance (R2), mean absolute error (MAE) and mean absolute percentage error (MAPE) values of 1.786 MPa, 0.997, 1.329 MPa and 4.44 %. Both experimental and predicted values showed a generally increasing % reduction of compressive strengths with increasing % plastic fine aggregate.
 The model showed reasonably low errors, reasonable accuracy and good generalization. ANN model could be used extensively in modeling of green concrete, with partial waste plastic fine aggregate. The study recommend ANNs models application as possible alternative for green concrete trial mix design. Sustainable techniques such as low-cost superplasticizers from recycled material and cost-effective technologies to adequately sizing and shaping plastic for fine aggregate application should be encouraged, so as to enhance strength of concrete with partial plastic aggregates.

Negative Correlation Hidden Layer for the Extreme Learning Machine

Extreme Learning Machine (ELM) algorithms have achieved unprecedented performance in supervised machine learning tasks. However, the preconfiguration of the nodes in the hidden layer in ELM models through randomness does not always lead to a suitable transformation of the original features. Consequently, the performance of these models relies on broad exploration of these feature mappings, generally using a large number of nodes in the hidden layer. In this paper, a novel ELM architecture is presented, called Negative Correlation Hidden Layer ELM (NCHL-ELM), based on the Negative Correlation Learning (NCL) framework. This model incorporates a parameter into each node in the original ELM hidden layer, and these parameters are optimized by reducing the error in the training set and promoting the diversity among them in order to improve the generalization results. Mathematically, the ELM minimization problem is perturbed by a penalty term, which represents a measure of diversity among the parameters. A variety of regression and classification benchmark datasets have been selected in order to compare NCHL-ELM with other state-of-the-art ELM models. Statistical tests indicate the superiority of our method in both regression and classification problems.

Real‑time COVID-19 diagnosis from X-Ray images using deep CNN and extreme learning machines stabilized by chimp optimization algorithm

Real-time detection of COVID-19 using radiological images has gained priority due to the increasing demand for fast diagnosis of COVID-19 cases. This paper introduces a novel two-phase approach for classifying chest X-ray images. Deep Learning (DL) methods fail to cover these aspects since training and fine-tuning the model's parameters consume much time. In this approach, the first phase comes to train a deep CNN working as a feature extractor, and the second phase comes to use Extreme Learning Machines (ELMs) for real-time detection. The main drawback of ELMs is to meet the need of a large number of hidden-layer nodes to gain a reliable and accurate detector in applying image processing since the detective performance remarkably depends on the setting of initial weights and biases. Therefore, this paper uses Chimp Optimization Algorithm (ChOA) to improve results and increase the reliability of the network while maintaining real-time capability. The designed detector is to be benchmarked on the COVID-Xray-5k and COVIDetectioNet datasets, and the results are verified by comparing it with the classic DCNN, Genetic Algorithm optimized ELM (GA-ELM), Cuckoo Search optimized ELM (CS-ELM), and Whale Optimization Algorithm optimized ELM (WOA-ELM). The proposed approach outperforms other comparative benchmarks with 98.25 % and 99.11 % as ultimate accuracy on the COVID-Xray-5k and COVIDetectioNet datasets, respectively, and it led relative error to reduce as the amount of 1.75 % and 1.01 % as compared to a convolutional CNN. More importantly, the time needed for training deep ChOA-ELM is only 0.9474 milliseconds, and the overall testing time for 3100 images is 2.937 s.

Spatial distribution of particulate matter 2.5 released from surface fuel combustion of Pinus koraiensis – A laboratory simulation study

High concentration particulate matter 2.5 released from forest fires, in addition to direct burns and asphyxia, PM2.5 is one of the main pollutants which threaten the safety of forest fire fighter. Therefore, to assess spatial distribution of PM2.5, a simulation study was conducted. Fuel beds with different moisture contents and loads were constructed. 144 times burning experiments were carried out under different wind speeds by using wind tunnel device. PM2.5 particles at different spatial points were collected and calculated. The results show that, in the two of three variables interaction between wind speed, fuel load, and, except fuel moisture content, wind speed and fuel load are positively correlated with the PM2.5 concentrations. From PM2.5 concentration which collected at each point in the horizontal and vertical directions, the overall trend is that PM2.5 concentration increases along the horizontal downwind direction (C and D higer than A and B) and the vertical upward direction (A and C higer than B and D) Based on BP neural network, the spatial distribution model of PM2.5 concentration with single hidden layer was established. The prediction accuracy of modeling samples and validation samples is balanced when hidden layer node is 5. This study will help to make reference for PM2.5 occupational exposure standards, forest fire smoke management and forest fire management in China.

Prediction of System-Level Energy Harvesting Characteristics of a Thermoelectric Generator Operating in a Diesel Engine Using Artificial Neural Networks

This study evaluated the potential of artificial neural networks (ANNs) to predict the system-level performance of a thermoelectric generator (TEG), whose performance depends on various variables including engine load, engine rotation speed, and external load resistance. Therefore, a Python code was developed to determine an optimal ANN structure by tracking the training/prediction errors of the ANN as a function of the number of hidden layers and nodes of hidden layers. The optimal ANN was trained using 484 output current (I)–load resistance (R) datasets obtained under three different engine rotation speeds and five different engine loads. The prediction accuracy of the ANN was validated by comparing 88 I–R datasets reproduced by the ANN using experimental data that were not used for training. In the validation procedure, differences of only 3.49% and 2.59% were observed in the experimental and ANN-predicted output power obtained for the 1000 rpm–0.8 MPa brake mean effective pressure (BMEP) and 1500 rpm–0.4 MPa BMEP scenarios, respectively. The exhaust gas flow characteristics were used for training and validation to predict the pumping loss caused by the installation of the TEG in the middle of the exhaust tailpipe with high accuracy. The results demonstrated that the ANN effectively reproduced datasets to fill the gaps between the discretized experimental results for all the experimental scenarios without any noticeable overfitting and underfitting. The net power gain obtained by the ANN exhibited a clear peak point for the engine rotation speed of 2000 rpm, which is difficult to obtain using experimental data.

Qualitative identification of waste textiles based on near-infrared spectroscopy and the back propagation artificial neural network

Hand sorting for different types of waste textiles is time-consuming, laborious and inaccurate. The non-destructive and efficient identification of fibers in waste fabrics is of great significance to the reuse of textile materials. In this paper, 593 samples were selected as the research objects, including polyester, cotton, wool, viscose, nylon, silk, acrylic, polyester/nylon, polyester/cotton, polyester/wool and silk/cotton waste textiles. The near-infrared spectrum of each sample was obtained by a portable near-infrared spectrometer, and the influence of environmental humidity and fabric thickness on the near-infrared spectrum of the sample was discussed to obtain the best test conditions. On this basis, the back propagation artificial neural network (BP-ANN) was applied to the qualitative classification of waste textiles to complete the automatic identification of fabric components in the sorting process. Firstly, a standard sample set was established by waveform clipping and normalization, and a BP-ANN deep web suitable for near-infrared spectroscopy was established. Then the BP network was trained according to the input near-infrared spectrum data of known sample categories and the classification results of the preset 11 types of labels, and the weights and thresholds of each layer were adjusted in the repeated training process. Finally, a 1500 × 100 × 11 network structure was established when the network error was the smallest, and the number of corresponding hidden layer nodes was 100. When the number of training steps was 500, the sum of squared errors reached 0.001, and the model recognition effect was the best. Meanwhile, the validity of the model was verified by inspecting additional 299 samples outside the model, and the recognition accuracy rate of the established model also exceeded 99%, which verified the effectiveness of the model. These results show that this near-infrared qualitative analysis model can more accurately classify and identify waste textiles, especially polyester waste textiles. In addition, it provides a new idea for the recycling and reuse of waste textiles for enterprises.

Research on gear fault diagnosis based on feature fusion optimization and improved two hidden layer extreme learning machine

Aiming at gear fault diagnosis in rotating machinery, a novel state assessment method based on feature fusion optimization and improved two-hidden-layer extreme learning machine (ITELM) was proposed in order to improve the fault diagnosis accuracy. Firstly, variational mode decomposition (VMD) and wavelet packet (WP) are employed to decompose the vibration signal, and statistical parameter features are extracted. Then, ReliefF algorithm is adopted to manage the extracted features and obtain the optimal feature subset. Finally, in order to avoid the network structure limitation of the existing two-hidden layer extreme learning machine with same hidden-layer node number, ITELM with different hidden-layer node number is proposed to evaluate the gear running states. Meanwhile, the empirical formula and fuzzy logic reasoning methods are developed in order to effectively determine the number of nodes for first and second hidden layers, respectively. The experimental results from different perspectives verified the effectiveness of the proposed method.

A Power Transformer Fault Diagnosis Method Based on Random Vector Functional-Link Neural Network

The random vector functional link (RVFL) network is suitable for solving nonlinear problems from transformer fault symptoms and different fault types due to its simple structure and strong generalization ability. However, the RVFL network has a disadvantage in that the network structure, and parameters are basically determined by experiences. In this paper, we proposed a method to improve the RVFL neural network algorithm by introducing the concept of hidden node sensitivity, classify each hidden layer node, and remove nodes with low sensitivity. The simplified network structure could avoid interfering nodes and improve the global search capability. The five characteristic gases produced by transformer faults are divided into two groups. A fault diagnosis model of three layers with four classifiers was built. We also investigated the effects of the number of hidden nodes and scale factors on RVFL network learning ability. Simulation results show that the number of implicit layer nodes has a large impact on the network model when the number of input dimensions is small. The network requires a higher number of implicit layer neurons and a smaller threshold range. The size of the scale factor has significant influence on the network model with larger input dimension. This paper describes the theoretical basis for parameter selection in RVFL neural networks. The theoretical basis for the selection of the number of hidden nodes, and the scale factor is derived. The importance of parameter selection for the improvement of diagnostic accuracy is verified through simulation experiments in transformer fault diagnosis.

Synergy evaluation model of container multimodal transport based on BP neural network

With the rapid development of economic globalization, the trade of various countries has become increasingly close and thus the rapid growth of container transportation business. The rational choice of transportation routes and intermodal transportation methods has become a hot topic in this field. In order to improve the efficiency of intermodal transportation and reduce the cost of intermodal transportation, this paper studies the evaluation of the synergy effect of container multimodal transportation based on the BP neural network algorithm. First, it is determined that the multi-attribute decision-making method is used to comprehensively evaluate the synergy effect of container multimodal transportation, and the data are normalized. Select the appropriate hidden layer nodes, use Matlab to determine the learning rate, select the logsig function for the network transfer function, select Traingdx as the training function, and establish a container multimodal transport synergy evaluation model based on BP neural network. Secondly, the model is solved, and the problem is transformed into finding the shortest route from Q to P without exceeding the cost and time constraints. Then carry on the simulation experiment, use Matlab to solve the transportation time and total cost between cities. Experimental data show that when the number of iterations is 800, the algorithm begins to converge; at the fifth training, the error between the actual output and the expected output is only 0.0004; in route B, the time of multimodal transportation is 48.18% less than that of single transportation, and the cost is saved by 50.02%. This shows that the container multimodal transport synergy evaluation model based on BP neural network can accurately evaluate the effects, and container multimodal transport can indeed improve transportation efficiency and save costs.

Prediction of Bending Force in the Hot Strip Rolling Process Using Multilayer Extreme Learning Machine

In the hot strip rolling process, accurate prediction of bending force can improve the control accuracy of the strip flatness and further improve the quality of the strip. In this paper, based on the production data of 1300 pieces of strip collected from a hot rolling factory, a series of bending force prediction models based on an extreme learning machine (ELM) are proposed. To acquire the optimal model, the parameter settings of the models were investigated, including hidden layer nodes, activation function, population size, crossover probability, and hidden layer structure. Four models are established, one hidden layer ELM model, an optimized ELM model (GAELM) by genetic algorithm (GA), an optimized ELM model (SGELM) by hybrid simulated annealing (SA) and GA, and two-hidden layer optimized ELM model (SGITELM) optimized by SA and GA. The prediction performance is evaluated from the mean absolute error (MAE), root-mean-squared error (RMSE), and mean absolute percentage error (MAPE). The results show that the SGITELM has the highest prediction accuracy in the four models. The RMSE of the proposed SGITELM is 11.2678 kN, and 98.72% of the prediction data have an absolute error of less than 25 kN. This indicates that the proposed SGITELM with strong learning ability and generalization performance can be well applied to hot rolling production.

Prediction of pharmacological activities from chemical structures with graph convolutional neural networks

Many therapeutic drugs are compounds that can be represented by simple chemical structures, which contain important determinants of affinity at the site of action. Recently, graph convolutional neural network (GCN) models have exhibited excellent results in classifying the activity of such compounds. For models that make quantitative predictions of activity, more complex information has been utilized, such as the three-dimensional structures of compounds and the amino acid sequences of their respective target proteins. As another approach, we hypothesized that if sufficient experimental data were available and there were enough nodes in hidden layers, a simple compound representation would quantitatively predict activity with satisfactory accuracy. In this study, we report that GCN models constructed solely from the two-dimensional structural information of compounds demonstrated a high degree of activity predictability against 127 diverse targets from the ChEMBL database. Using the information entropy as a metric, we also show that the structural diversity had less effect on the prediction performance. Finally, we report that virtual screening using the constructed model identified a new serotonin transporter inhibitor with activity comparable to that of a marketed drug in vitro and exhibited antidepressant effects in behavioural studies.

Comparison between the backpropagation and single exponential smoothing method in sugar production forecasting case

Sugar is a strategic commodity, because it is one of the basic needs consumed by the community. The decline in sugar production causes turmoil both economically and politically. So companies must be able to balance sugar production in accordance with market demand. Because, if the sugar production exceeds market demand, it will cause accumulation and expenditure costs that exceed the limit. Conversely, if the sugar production is too little, then the company is unable to meet demand. This study was comparing the Single Exponential Smoothing method and the Backpropagation method in predicting the amount of sugar production. The data used was based on production in 2013-2017, with predictions every month starting in the fifth month. The alpha parameter in Single Exponential Smoothing was 0.84, while in Backpropagation used 5 input nodes and 5 hidden layer nodes. From the results of tests that have been carried out, it can be seen that the Backpropagation method produces the smallest MSE and MAPE, amounting to 57187817.49 and MAPE of 1.012% compared to the Single Exponential Smoothing method with MSE of 83602989.43 and MAPE of 1.46%.

Short-Term Wind Power Prediction Based On Particle Swarm Optimization-Extreme Learning Machine Model Combined With Adaboost Algorithm

As the proportion of wind power in the world's electricity generation increases, improving wind power prediction accuracy is vital for making full use of wind energy and ensuring the safe and stable operation of the power grid. Given the uncertainty and volatility of wind power and the weak generalization ability of the current wind power prediction models, we propose a wind power prediction model that combines Adaboost algorithm with extreme learning machine optimized by particle swarm optimization (PSO-ELM). First, particle swarm optimization is used to optimize the initial thresholds and input weights of the ELM to obtain the PSO-ELM basic prediction model. Then, combined with the Adaboost algorithm, a series of PSO-ELM weak predictors with input weights and thresholds optimized by PSO and containing different hidden layer nodes are composed. Finally, each weak predictor is weighted and fused into a strong prediction model of wind power, and the final prediction results are output. In this paper, the Adaboost-PSO-ELM model is verified by a wind turbine's measured data in Turkey. The prediction indicators are compared with the current wind power prediction methods including optimized neural networks and ensemble learning models. The results show that the Adaboost-PSO-ELM wind power prediction model has higher accuracy and better generalization ability.

Optimal Variational Mode Decomposition and Integrated Extreme Learning Machine for Network Traffic Prediction

Network traffic prediction plays a vital role in effective network management, load evaluation and security warning. Extreme learning machine has the advantages of fast convergence speed and strong generalization ability. Also, it does not easily fall into local optima. The evolutionary algorithm can be used to optimize the number of its hidden layer nodes. However, most of the existing evolutionary algorithms have some adjustable parameters which depend on subjective experience or prior knowledge. Hence, this can affect the model prediction accuracy. Given this context, this paper proposes a network traffic prediction mechanism based on optimized Variational Mode Decomposition (VMD) and Integrated Extreme Learning Machine (ELM). A Scalable Artificial Bee Colony (SABC) algorithm which has fewer adjustable parameters and can thus guarantee the accuracy and stability of the prediction mechanism is also proposed. It can be used in the optimization selection of VMD, Phase Space Reconstruction (PSR) and ELM to achieve higher prediction performance. Finally, we utilize Mackey-Glass, Lorenz chaotic time series of recognized benchmark and a WIDE backbone actual network traffic data to prove the validity of the proposed network traffic prediction mechanism.

The Improved ELM Algorithms Optimized by Bionic WOA for EEG Classification of Brain Computer Interface

The breakthrough of electroencephalogram (EEG) signal classification of brain computer interface (BCI) will set off another technological revolution of human computer interaction technology. Because the collected EEG is a type of nonstationary signal with strong randomness, effective feature extraction and data mining techniques are urgently required for EEG classification of BCI. In this paper, the new bionic whale optimization algorithms (WOA) are proposed to promote the improved extreme learning machine (ELM) algorithms for EEG classification of BCI. Two improved WOA-ELM algorithms are designed to compensate for the deficiency of random weight initialization for basic ELM. Firstly, the top several best individuals are selected and voted to make decisions to avoid misjudgment on the best individual. Secondly, the initial connection weights and bias between the input layer nodes and hidden layer nodes are optimized by WOA through bubble-net attacking strategy (BNAS) and shrinking encircling mechanism (SEM), and different regularization mechanisms are introduced in different layers to generate appropriate sparse weight matrix to promote the generalization performance of the algorithm.As shown in the contrast results, the average accuracy of the proposed method can reach 93.67%, which is better than other methods on BCI dataset.

Application of Mathematical Economic Model in Financial System in Manufacturing Industry

In the field of economic research, most of the sample data is not obtained based on controllable experiments but generated during the normal operation of the economic system. Therefore, the change of an economic variable is usually not caused by a single change of a cause variable. It is the result of a combination of multiple factors. Therefore, it is necessary to study the application of mathematical intelligent computing in computer intelligent manufacturing system. The purpose of this paper is to explore the application of mathematical intelligent computing in computer intelligent manufacturing system. For this reason, this paper uses the furnace temperature control model to carry out simulation experiment. In this simulation experiment, three algorithms of mathematical intelligent computing are mainly used, including BPES intelligent computing method, genetic algorithm, and MARS algorithm. The research results show that the superparameter optimization based on MARS has high efficiency, and the best result, the worst result, the average result, the variance, and the average time of multiple independent runs are controlled below 0.03 s. In this experiment, when the hidden layer node is 9, the prediction error value is the smallest, and the model simulation curve is basically consistent with the measured curve trend. In the simulation experiment of this paper, these three algorithms have shown good results in their respective links.

Research on Plant Leaf Recognition Based on BP Neural Network

Aiming at the problems of single plant identification feature, complex algorithm and unsatisfactory recognition rate, this research proposes to use BP algorithm to establish a plant leaf identification model. This research pre-processed the image of plant leaves, and extracted 14 feature variables of the image, including colour features, shape features and texture features. Using SPSS 21.0 software to perform principal component analysis on the extracted 14 feature variables, the analysis obtained 6 the main ingredient. Using the recognition model to carry out a simulation experiment, the experimental results show that when the number of hidden layers of the model is 1, the number of hidden layer nodes is 6, the accuracy of model recognition is better, and the average accuracy of model recognition is 95.56%. From the analysis of model performance, the training accuracy and recognition accuracy of the model did not fluctuate greatly, and the graph changes relatively smoothly; when learning times is 400 times, the training and testing accuracy of the model reached more than 95%. It shows that the model has high accuracy and strong generalization ability, which provides reference for plant leaf recognition research.