Number Of Neurons In Layer Research Articles

A multi-energy load prediction of a building using the multi-layer perceptron neural network method with different optimization algorithms

Since cooling and heating loads are recognized as key characteristics for evaluating the energy efficiency of buildings, it appears indisputable that they must be predicted and analyzed for residential structures. Accordingly, the multi-layer perceptron neural network is applied for predicting the heating and cooling loads using the experimental dataset. The used dataset is obtained by monitoring the impact of the building's dimensions on energy consumption. To optimize the training process of the multi-layer perceptron neural network, several optimizers are employed. Besides, different statistical performance indicators are considered to see which selected optimizer outperforms the rest in terms of accuracy and authenticity. The obtained results emphasize the remarkable performance of adaptive chaotic grey wolf optimization, which can be used to train the multi-layer perceptron neural network and forecast the buildings’ energy consumption with the highest accuracy. According to the obtained results, the hybrid multi-layer perceptron neural network- adaptive chaotic grey wolf optimization method demonstrates the best performance. The optimum number of neurons in the hidden layer is obtained to be 15. Also, based on the statistical performance indicators, the selected method reveals an R2 of 0.9123 and 0.9419 for cooling and heating loads, respectively.

Optimization of the structural complexity of artificial neural network for hardware-driven neuromorphic computing application

This work focuses on the optimization of the structural complexity of a single-layer feedforward neural network (SLFN) for neuromorphic hardware implementation. The singular value decomposition (SVD) method is used for the determination of the effective number of neurons in the hidden layer for Modified National Institute of Standards and Technology (MNIST) dataset classification. The proposed method is also verified on a SLFN using weights derived from a synaptic transistor device. The effectiveness of this methodology in estimating the reduced number of neurons in the hidden layer makes this method highly useful in optimizing complex neural network architectures for their hardware realization.

Capturing the latent space of an Autoencoder for multi-omics integration and cancer subtyping

Background and Objective:The motivation behind cancer subtyping is to identify subgroups of cancer patients with distinguishable phenotypes of clinical importance. It can assist in advancement of subtype-targeted based treatments. Subtype identification is a complicated task, therefore requires multi-omics data integration to identify the precise patients’ subgroup. Over the years, several computational attempts have been made to identify the cancer subtypes accurately using integrative multi-omics analysis. Some studies have used Autoencoders (AE) to capture multi-omics feature integration in lower dimensions for identifying subtypes in specific types of cancer. However, capturing the highly informative latent space by learning the deep architectures of AE to attain a satisfactory generalized performance is required. Therefore, in this study, a novel AE-assisted cancer subtyping framework is presented that utilizes the compressed latent space of a Sparse AE neural network for multi-omics clustering. Methods:The proposed framework first performs a supervised feature selection based on the survival status of the patients. The selected features from each of the omic data are passed to the AE. The information embedded in the latent space of the trained AE neural networks are then used for cancer subtyping using Spectral clustering. The AE architecture designed in this study exhaustively searches the best compression for multi-omics data by varying the number of neurons in the hidden layers and penalizing activations within the layers. Results and Conclusion:The proposed framework is applied to five different multi-omics cancer datasets taken from The Cancer Genome Atlas. It is observed that for getting a robust information bottleneck, a compression of 10–20% of the input features along with an L1 regularization penalty of 0.01 or 0.001 performs well for most of the cancer datasets. Clustering performed on this latent representation generates clusters with better silhouette scores and significantly varying survival patterns. For further biological assessment, differential expression analysis is performed between the identified subtypes of Glioblastoma multiforme (GBM), followed by enrichment analysis of the differentially expressed biomarkers. Several pathways and disease ontology terms coherent to GBM are found to be significantly associated. Varying responses of the identified GBM subtypes towards the drug Temozolomide is also tested to demonstrate its clinical importance. Hence, the study shows that AE-assisted multi-omics integration can be used for the prediction of clinically significant cancer subtypes.

Experimental investigation on end winding thermal management with oil spray in electric vehicles

To achieve a carbon-neutral sustainable future, the transportation sector is experiencing quick electrification, as more and more vehicles are driven by electric motors. Electric motors are designing more and more powerful and thus pose a great increasing challenge for thermal management. In this paper, an innovative experimental study on oil spray for end winding thermal management were presented. The experimental results indicated that the oil spray cooling significantly decrease the temperature of the winding. With higher oil temperature and flow rate, better cooling performance and better temperature uniformity were achieved both for teeth and slot winding. Meanwhile, a heat transfer performance prediction model based on a fully connected artificial neural network (ANN) was developed to correlate oil spray. The model used Reynolds number and Prandtl number as inputs, and the number of neurons in hidden layer was optimized to 8. The model accurately predicts the heat transfer performance of oil spray with a correlation coefficient of 0.93. The paper is expected to be a good reference for the oil spray engineering application in e-motor cooling.

The Formula Study in Determining the Best Number of Neurons in Neural Network Backpropagation Architecture with Three Hidden Layers

The researchers conducted data simulation experiments, but they did so unstructured in determining the number of neurons in the hidden layer in the Artificial Neural Network Back-Propagation architecture. The researchers also used a general architecture consisting of one hidden layer. Researchers are still producing minimal research that discusses how to determine the number of neurons when using hidden layers. This article examines the results of experiments by conducting training and testing data using seven recommended formulas including the Hecht-Nelson, Marchandani-Cao, Lawrence & Fredrickson, Berry-Linoff, Boger-Guterman, JingTao-Chew, and Lawrence & Fredrickson modifications. We use rainfall data and temperature data with a 10-day type for the last 10 years (2012-2021) sourced from Lombok International Airport Station, Indonesia. The training and testing data used showed the results that in determining the number of neurons on the hidden-1 screen, it was more appropriate to use the Hecht-Nelson formula and the Lawrence & Fredricson formula which is more suitable for use in the 2nd & 3rd hidden layer. The resulting research was able to provide an accuracy rate of up to 97.79% (temperature data) and 99.94% (rainfall data) with an architecture of 36-73-37-19-1.

Research on State Trend Prediction of Hydraulic Turbine Units Based on Mult-dimensional and Multi-condition Data

Aiming at the problem that it is difficult to accurately predict the flexible operation conditions and state trend of hydraulic turbine, in this paper a cascade model of BP-LSTM classification prediction is proposed, which can identify the working conditions of existing fusion data, and then predict the measuring points of different working conditions.<BR> Based on the pressure parameters of hydraulic turbine units, the improved BP neural network is used to determine the operation conditions of hydraulic turbine units, and the classified data is redivided to establish the multivariate LSTM prediction model. By optimizing the parameters of the multivariate LSTM prediction model, such as the structure, the number of network layers and the number of hidden layer neurons, finally established the cascade model of BP-LSTM classification prediction of time series of hydraulic turbine units.<BR> Through experimental verification and analysis, BP-LSTM classification prediction model can predict the operation trend of measuring points under different working conditions after classification. Compared with other models, BP-LSTM model has higher prediction accuracy and better effect. The cascade model of BP-LSTM classification prediction of time series provides a model basis for the research of predictive control of hydraulic turbine units.

An integrated framework of gated recurrent unit based on improved sine cosine algorithm for photovoltaic power forecasting

Accurate prediction of photovoltaic power is of great significance to the storage and utilization of solar power. In this research, a deep learning model for photovoltaic power prediction based on gated recurrent unit network (GRU), improved sine cosine algorithm (ISCA), and complete ensemble empirical mode decomposition (CEEMD) is proposed. Firstly, CEEMD is used to decompose the original photovoltaic data into several intrinsic mode function (IMF) components and one residual. Secondly, each sub-pattern after decomposition is processed by partial least-squares analysis (PLS). Third, the nonlinear strategy is used to improve SCA, and the Hill-climbing strategy is added to the local search part to improve the performance of the algorithm. Fourth, each sub-pattern is predicted by GRU, then the learning rate and the number of hidden layer neurons of GRU are optimized by the ISCA. Finally, the predicted results of each sub-model are combined to generate the final prediction results. In this study, the proposed model is applied to four photovoltaic power data sets, and different experimental comparison models are established. The experimental results show that the CEEMD-PLS-ISCA-GRU model in this study can obtain good prediction results in all data sets.

Forecasting of Reactive Power Consumption with the Use of Artificial Neural Networks

Many electricity customers, particularly in the area of small and medium-sized enterprises, are characterized by a system of 3-phase unbalanced current consumption and the polarization of power factor in the individual phases. Strong variation of power factor in the different phases (often with both inductive and capacitive nature) causes the inability to install classic–a 3-phase reactive power compensators. Therefore, energy consumers are exposed to higher costs for the crossing of the contractual power factor. This paper describes the problem of reactive power forecasting with the use of artificial neural networks. For calculation, the Nonlinear Autoregressive (NAR) neural network was used for different input vector sizes and different numbers of neurons in the hidden layer for foreseeing reactive power generation. Results of simulations compared to real measurements confirm that it is possible to forecast the reactive power course, useful for optimal planning of reactive power compensation strategies.

Application analysis of digital fund prediction model based on neural network

SummaryWith the development of the economy, digital funds have developed rapidly, and traditional big data processing technologies have been unable to effectively solve the update of digital funds information. This research uses neural network to study the digital fund forecasting model based on big data technology and deep learning algorithm. First, on the basis of the feedforward neural network model, the network structure and network parameters are optimized to effectively improve the algorithm's ability to learn the characteristics of digital capital information. Through the first‐order approximation, the number of iterations in the model parameter update is effectively reduced, and the computational efficiency of the algorithm is improved. By improving the model network structure, increasing the number of neurons in the hidden layer, and improving the dynamic change learning ability of the algorithm, the dynamic characteristics of the incremental file information can be effectively obtained. Through the experimental simulation calculation, it can be seen that compared with the traditional algorithm, the algorithm proposed in this study has certain advantages in terms of convergence speed, convergence ability and solution error, and the algorithm has good stability and strong inheritance, and is suitable for digital capital forecasting aspects of research.

Quality prediction and control of thin-walled shell injection molding based on GWO-PSO, ACO-BP, and NSGA-II

Abstract ECG recorders are precision medical devices, but their thin-walled shells are susceptible to warpage and shrinkage during injection molding production due to the injection molding process, which greatly shortens their service life. To address this problem, a multiobjective optimization method for injection molding process parameters based on a combination of a BP neural network model optimized by an ant colony algorithm (ACO-BP) and an improved non-dominated sorting genetic algorithm (NSGA-II) is proposed. The study takes the warpage deformation amount and volume shrinkage rate of the plastic part as the optimization objectives, and the melt temperature, mold temperature, injection pressure, holding pressure, holding time, and cooling time as the design variables. However, for BP neural networks, it is crucial to choose an appropriate number of hidden layer neurons, so the particle swarm algorithm combined with the grey wolf algorithm (GWO-PSO) is used to solve for the optimal number of hidden layer neurons. Firstly, the number of hidden layer neurons of the BP network model was solved based on the samples obtained from the Box–Behnken experimental design and the GWO-PSO algorithm, and the ACO-BP algorithm was used to build the prediction models for warpage and volume shrinkage, respectively, and then combined with NSGA-II for global optimisation. The pareto optimal solution set was subjected to CRITIC analysis and the optimal process parameters were finally obtained, with a minimum warpage of 0.3293 mm and minimum volume shrinkage of 4.993%, a reduction of 8.93 and 6.95% respectively compared to the pre-optimisation period. At the same time, injection molding tests were carried out on the optimum process parameters, and it was found that the molding quality of the plastic parts was better and met the actual production requirements through measurement. The research in this paper provides a theoretical basis for further improving the quality defects of the thin-walled injection molded parts.

Hybrid Models for Solving the Colebrook–White Equation Using Artificial Neural Networks

This study proposes hybrid models to solve the Colebrook–White equation by combining explicit equations available in the literature to solve the Colebrook–White equation with an error function. The hybrid model is in the form of fH=fo−eA. fH is the friction factor value f predicted by the hybrid model, fo is the value of f calculated using several explicit formulas for the Colebrook–White equation, and eA is the error function determined using the neural network procedures. The hybrid equation consists of a series of hyperbolic tangent functions whose number corresponds to the number of neurons in the hidden layer. The simulation results showed that the hybrid models using five hyperbolic tangent functions could produce reasonable predictions of friction factors, with the maximum absolute relative error (MAXRE) around one tenth, or ten times lower than that produced by the corresponding existing formula. The simplified hybrid models are also given using four and three tangent hyperbolic functions. These simplified models still provide accurate results with MAXRE of less than 0.1%.

DEVELOPMENT AND APPLICATION OF NEURAL NETWORKS TECHNOLOGY FOR DIAGNOSTICS OF THE TECHNICAL CONDITION OF AUTOMOBILE AND TRACTOR ENGINES

The article proposes the use of neural network technology to solve the problems of diagnosing the automobile and tractor engines technical conditions. This allow to work with real data obtained for an individual and reference (average) engine, as well as with data calculated using the mathematical model, based on the comparison of which researcher can take informed decisions about the nature and location of a particular defect. This improves the functional stability of the wheeled vehicle and improves the performance of its intelligent on-board systems. It has been proven that in order to reduce the redundancy of a neural network, it is necessary to reduce the number of neurons in the hidden layer for a given level of network training error. For diagnosing the automobile or tractor engine, the input signals are voltages received in an artificial neural network from sensors that are standard in the engine, and additionally indicate the technical conditions.

Stability Analysis of Geotechnical Landslide Based on GA-BP Neural Network Model.

Rock and soil landslides, a regular geological disaster in engineering construction, endanger national property and, in severe circumstances, result in a huge number of casualties. A set of methods for landslide stability analysis and prediction has been established, with the academic idea of “geological process mechanism analysis-quantitative evaluation” at its core, combined with detailed field investigation of geological hazards, forming a relatively complete technical route for research on landslide stability analysis. The work of this paper can be summarized as follows: (1) Introduce the research status of geotechnical landslide stability at home and abroad and the current development trend of neural network. (2) Through the collected sample database, take the training function and the number of hidden layer neurons as variables to optimize the BP neural network, and combine the optimized BP neural network with the genetic algorithm to construct the GA-BP neural network. (3) The stability coefficients of the BP neural network, the genetic algorithm based back propagation neural network (GA-BPNN), and the limit equilibrium technique are analyzed and compared. The findings imply that landslide stability can be assessed using neural networks. GA-BPNN is a viable alternative to back propagation neural network (BPNN). The algorithm is more accurate, has a faster convergence rate, and is more stable.

Neural network estimations of annealed and non-annealed Schottky diode characteristics at wide temperatures range

In this study, Artificial Neural Network (ANN) model has been proposed to characterize the annealed and the non-annealed Schottky diode from experimental data. The experimental current values of Ni/n-type 6H–SiC Schottky diode for the voltages applied to the diode terminal starting from 80 K with 20 K steps up to 500 K temperature were measured for both non-annealed and annealed Schottky diodes. The applied voltage has been varied starting from -2 V with 10 mV steps up to +2 V for each temperature value. The modeling performance has been assessed according to the varying number of neurons in the hidden layer, starting from 5 to 50 neurons, thereafter the optimum number of neurons has been obtained for both annealed and non-annealed ANN models. The minimum Root Mean Square Error (RMSE) and Mean Absolute Error (MAE) indices values for both annealed and non-annealed diodes have been obtained with 40 neurons for both the training and test phase.

Modified binary salp swarm algorithm in EEG signal classification for epilepsy seizure detection

Epilepsy is a brain disorder characterized by sudden seizures, periodic abnormal and inappropriate behaviour, and an altered state of consciousness. The visual diagnosis of epilepsy using electroencephalogram (EEG) signals is challenging, which led to the development of machine learning methods to automate this task. With the help of machine learning techniques optimized for epilepsy, this study aims to diagnose epilepsy disorders and related seizures with high accuracy. In the first step, the proposed multilevel method applies Discrete Wavelet Transform (DWT) to decompose the EEG signal into sub-band frequency levels. Next, the algorithm uses the Modified Binary Salp Swarm Algorithm (MBSSA), a population-based strategy, to extract time-domain features. The proposed method uses the Levenberg-Marquardt (LM) backpropagation classification model as a Feed-Forward Neural Network (FFNN). The MBSSA also optimally determine the type of WT (DWT or Double Density DWT (DDT)), the number of decomposed levels in WT, the best-fitted mother wavelet and the number of neurons in the hidden layer of FFNN, preventing manual and time-consuming calculation. Evaluation procedures compare the proposed method to other state-of-the-art methods and verify its superiority by achieving the highest and average accuracy of %99.45 and %98.46, respectively.

Traffic Sign Recognition and Classification Using Deep Convolution Neural Network

Traffic signs classification is that the process of identifying which class a traffic sign belongs to. There are several differing kinds of traffic signs like speed limits, no entry, turn left or right, etc. A neural network model based on deep learning is utilized to explore the traffic sign recognition (TSR) and expand the application of deep intelligent learning technology in the field of virtual reality (VR) image recognition, thereby assessing the road traffic safety risks and promoting the construction of intelligent transportation networks Firstly, a dual-path deep CNN (TDCNN) TSR model is constructed supported CNN, and recognition accuracy is calculated to research the training results of the model. Secondly, we assess the road traffic safety risks, and also the prediction and evaluation effects. Finally, the changes in safety risks of road traffic accidents are analyzed based on the two key influencing factors of the number of road intersections and the speed of vehicles traveling. The results show that the learning rate of the network model and the number of hidden neurons in the fully-connected layer directly affect the training results, and there are differences in the choices between the early and late training periods. It instantly assists drivers or automatic driving systems in detecting and recognizing traffic signs effectively. Key Words: Deep Convolutional Neural Network CNN, TensorFlow, Image Processing, ReLU Activation Function.

Modeling of artificial neural networks for silicon prediction in the cast iron production process

The main way to produce cast iron is in the blast furnace. In the production of hot metal, the control of silicon is important. Alumina and silica react chemically with limestone and dolomite to form blast furnace slag. In this work, 12 artificial neural networks (ANNs) were modeled with different numbers of neurons in each hidden layer. The number of neurons varied between 10 and 200 neurons. ANNs were used to predict the silicon content of hot metal produced. The ANN with 30 neurons showed the best performance. In the test phase, the mathematical correlation was 97.5% and the mean square error (MSE) was 0.0006, and in the cross-validation phase, the mathematical correlation was 95.5% while the MSE was 0.00035.

Application of BP neural network in on-board charger fault diagnosis

With the accelerated growth of the number of electric vehicles, the frequent occurrence of electric vehicle charging safety accidents has brought some safety hazards to people. The fault diagnosis of on-board charger circuit has gradually become the focus of attention of scholars. This paper uses simulation software to conduct modeling simulation analysis, failure mode analysis and data acquisition for the on-board charger of electric vehicles, and establishes a fault diagnosis model based on BP neural network, which is applied to the fault diagnosis of on-board charger circuits. The optimal fault diagnosis model is obtained by continuously adjusting the number of neurons in the hidden layer of the BP neural network, and the open-circuit faults of diodes and MOSFETs are mainly analyzed experimentally. The simulation results show that the expected effect has been achieved in the fault diagnosis of the on-board charger circuit, and the established simple BP neural network model can be applied in practice. Some feasible solutions are provided for improving the charging safety of electric vehicles.

A Novel Recognition Method of Aging Stage of Transformer Oil-Paper Insulation Using Raman Spectroscopic Recurrence Plots

Timely identification of the aging stage of the transformer can effectively ensure the safe operation of the transformer and prevent the aging fault of the transformer. This work proposes the aging diagnosis of transformer using the Raman spectroscopic recurrence plot combined with learning vector quantization (LVQ) neural network. Oil-paper insulation samples at different aging stages are obtained by accelerated thermal aging tests. Acquisition of Raman spectra of insulating oil is from 800 to 3000 cm <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">−1</sup> and converts to the Raman spectroscopic recurrence plot. The gray-level cooccurrence matrix (GLCM) is used to extract the characteristic information of the Raman spectroscopic recurrence plot. The optimal number of neurons in competitive layer of LVQ neural network is determined by the experimental method. A transformer aging diagnosis model based on the Raman spectroscopic recurrence plot is established. The results show that the diagnostic accuracy of the proposed diagnostic model reached 95%. The Raman spectroscopic recurrence plot has more abundant aging “fingerprint” characteristic information and the proposed diagnostic method shows potential for monitoring oil-paper insulation.

Prognostication of scour around twin and three piers using efficient outlier robust extreme learning machine.

One of the most essential difficulties in the design and management of bridge piers is the estimation and modeling of scouring around the piers. The scour depth downstream of twin and three piers were simulated using a new outlier robust extreme learning machine (ORELM) model in this study. Furthermore, k-fold cross-validation with k = 4 was employed to validate the outcomes of numerical models. Four ORELM models with effective scouring parameters were first created to simulate scour depth. After then, the number of hidden layer neurons increased from two to thirty. The number of ideal hidden neurons was determined by examining the modeling results. The sigmoid activation function was also introduced as the best function. Furthermore, a sensitivity analysis was used to identify the superior model. The best model predicted scour depth as a function of the Froude number (Fr), the pier diameter to flow depth ratio (D/h), and the distance between the piers to flow depth ratio (d/h). The values of the objective function were accurately approximated by this model. As a result, using the ORELM model, the R2, scatter index, and Nash-Sutcliffe efficiency coefficient were calculated to be 0.953, 0.146, and 0.949, respectively. The most efficient parameters for simulating the scour depth were Fr and D/h, according to the modeling results. It is worth noting that nearly half of the superior model's simulated outputs had an inaccuracy of less than 10%. The superior model's performance has been underestimated, according to uncertainty analysis. After that, a simple and practical equation for calculating the scour depth was established for the superior model. Additionally, the influence of each input parameter on the objective function was assessed using a partial derivative sensitivity analysis.