Backpropagation Method Research Articles

Rainfall Prediction using Backpropagation Method and GIS for Disaster Mitigation Mapping

Frequent weather or climate changes, coupled with the topography of Cirebon and Majalengka which are mostly lowlands and close to the sea, raise concerns that flood disasters can occur if mitigation is not done early. This research aims to predict and map rainfall and runoff discharge as a flood mitigation effort in the Cirebon and Majalengka areas. The method used in this research is the Backpropagation Method, which resembles the way the neuron system works in the human brain to learn patterns, optimize weights and biases, and reduce the error rate (mean square error) so that the predicted value is closer to the actual value. The results showed that Cirebon was predicted to have the highest rainfall of 2471 mm in 2048, based on prediction data from BMKG Penggung Station, while Majalengka was predicted to have the highest rainfall of 2922 mm in 2048, based on prediction data from BMKG Kertajati Station. The largest runoff discharge in the Cirebon region occurred in Susukan District (51.96 km²) with a discharge of 68.648 m³/second (Q25), and in the Majalengka region occurred in Lemahsugih District (78.64 km²) with a discharge of 104.16 m³/second (Q25). The implications of this research include mitigation efforts that can be carried out, such as reducing development above drainage channels, reforestation, maintaining open areas, reducing land use change, conservation of rivers and catchments, construction of flood embankments, and issuance of related regulations.

IMPLEMENTATION OF BACKPROPAGATION AND HYBRID ARIMA-NN METHODS IN PREDICTING ACCURACY LEVELS OF RAINFALL IN MAKASSAR CITY

Hybrid ARIMA-NN is a combined approach of the ARIMA model used to capture linear patterns in time series data and Artificial Neural Networks (ANN) to handle non-linear and stochastic patterns. Using a gradient descent algorithm, backpropagation adjusts synaptic weights based on the error between the network's prediction and actual training data values. In this study, a comparison was made between the Backpropagation method and Hybrid ARIMA-NN in forecasting rainfall in Makassar City. Rainfall data in Makassar City uses data from the rainfall measuring station at the Paotere Maritime Meteorological Station in Makassar. The activation functions used are ReLU and Leaky ReLU with epoch parameters set at 350, and learning rates of 0.01, 0.001, 0.0001, and 0.00001. The two best methods selected for further evaluation are Backpropagation with architecture 12-32-16-8-1 and Hybrid ARIMA-NN (ARIMA [4,0,1]-NN 12-256-128-64-1). The ARIMA model (4,0,1) with AIC values of 1303.4 and RMSE 162,369 is the best compared to other models, which aligns with the advantages of backpropagation architecture. The results showed that the Backpropagation method excelled with an RMSE value of 137.320 or 0.1149, indicating high accuracy in forecasting changes in seasonal trends and patterns. Hybrid ARIMA-NN gives good results with RMSE 145.834, as residues contain better nonlinearity compared to ARIMA models (4,0,1), although it shows a slightly higher error rate compared to Backpropagation.

Estimation of simultaneous equation models by backpropagation method using stochastic gradient descent

Simultaneous equation model (SEM) is an econometric technique traditionally used in economics but with many applications in other sciences. This model allows the bidirectional relationship between variables and a simultaneous relationship between the equation set. There are many estimators used for solving an SEM. Two-steps least squares (2SLS), three-steps least squares (3SLS), indirect least squares (ILS), etc. are some of the most used of them. These estimators let us obtain a value of the coefficient of an SEM showing the relationship between the variables. There are different works to study and compare the estimators of an SEM comparing the error in the prediction of the data, the computational cost, etc. Some of these works study the estimators from different paradigms such as classical statistics, Bayesian statistics, non-linear regression models, etc. This work proposes to assume an SEM as a particular case of an artificial neural networks (ANN), considering the neurons of the ANN as the variables of the SEM and the weight of the connections of the neurons the coefficients of the SEM. Thus, backpropagation method using stochastic gradient descent (SGD) is proposed and studied as a new method to obtain the coefficient of an SEM.

A Novel Modeling Optimization Approach for a Seven-Channel Titania Ceramic Membrane in an Oily Wastewater Filtration System Based on Experimentation, Full Factorial Design, and Machine Learning.

This comprehensive study looks at how operational conditions affect the performance of a novel seven-channel titania ceramic ultrafiltration membrane for the treatment of produced water. A full factorial design experiment (23) was conducted to study the effect of the cross-flow operating factors on the membrane permeate flux decline and the overall permeate volume. Eleven experimental runs were performed for three important process operating variables: transmembrane pressure (TMP), crossflow velocity (CFV), and filtration time (FT). Steady final membrane fluxes and permeate volumes were recorded for each experimental run. Under the optimized conditions (1.5 bar, 1 m/s, and 2 h), the membrane performance index demonstrated an oil rejection rate of 99%, a flux of 297 L/m2·h (LMH), a 38% overall initial flux decline, and a total permeate volume of 8.14 L. The regression models used for the steady-state membrane permeate flux decline and overall permeate volume led to the highest goodness of fit to the experimental data with a correlation coefficient of 0.999. A Multiple Linear Regression method and an Artificial Neural Network approach were also employed to model the experimental membrane permeate flux decline and analyze the impact of the operating conditions on membrane performance. The predictions of the Gaussian regression and the Levenberg-Marquardt backpropagation method were validated with a determination coefficient of 99% and a Mean Square Error of 0.07.

MATHEMATICAL APPROACH OF THE BACKPROPAGATION METHOD FOR CONSTRUCTING ARTIFICIAL NEURAL NETWORKS

Backpropagation isthe core part of a neural network. This method is used to efficiently train a network using a chain rule that allows differentiation of complex functions. In other words, after each pass through the network, the backpropagation method performs a backward pass to adjust the model parameters, such as weights and biases. This article highlights the importance of using the backpropagation method from the point of view of mathematical formulas for neural networks. The importance of using the backpropaga-tion learning algorithm to calculate the gradient (gradient descent) and the need to use the activation function to minimize the loss function is mathematically described and calculated by formulas, and also proven by calculating the matrix products of vectors for each layer of parameters - weights and biases and applying complex differential equations.

Jaringan Saraf Tiruan (JST) Memprediksi Penyakit Rubella Menggunakan Metode Backpropagation

The development of Information Technology is now entering various sectors, including health, and is implemented in Bidadari Binjai Hospital. As a health institution that is committed to excellent service and quality, Bidadari Binjai Hospital needs to innovate technology. One health issue that requires attention is rubella, an airborne infectious disease that has the potential to cause serious disorders such as hearing loss, cataracts, speech delay, and heart failure in toddlers and children. The initial symptoms of rubella are often similar to other common diseases, so public understanding of these symptoms is very important for quick treatment. This research aims to develop an information technology-based system that is able to predict rubella using the backpropagation method. This method is expected to improve the accuracy of diagnosis and make it easier for people to recognize rubella symptoms early on. The proposed system aims to provide better diagnosis support at Bidadari Binjai Hospital, as well as increase public awareness and knowledge about rubella disease. From the research conducted, the results of the accuracy rate obtained when conducting a test program were selected according to the symptoms and the results obtained were rubella disease with a 100% accuracy rate.

Jaringan Saraf Tiruan (JST) Memprediksi Penjualan UMKM Kota Binjai dengan menggunakan Metode Backpropagation

The development of information technology that is currently developing serves to facilitate, accelerate, benefit and provide other alternatives for people who have businesses and have a big influence in the future. One of the things that is very influential is the sale of MSMEs. MSMEs are productive businesses owned by individuals or business entities that have met the criteria as micro businesses that have an important role in the economy because they provide employment, encourage local economic growth, and create innovation. MSMEs still face challenges such as limited access to financing, digital readiness, and marketing access that hinder the development of MSMEs. Therefore, it is necessary to take action to predict MSME sales in Binjai City using the backpropagation method so that later it can create new innovations and encourage community economic growth. Based on the process carried out using the backpropagation method, it can be seen that the value obtained has reached more than the predetermined target with a target value (t) of 0.26, learningrate 0.2, maximum epoch 10000 target error 0.01.

Artificial neural network prediction on free convection in a triangular chamber with a v-shaped corrugated bottom wall filled with nanofluid

Artificial neural network (ANN) prediction on magnetohydrodynamic (MHD) free convection for complex systems using nanofluid plays a significant role in understanding and optimizing the system’s behaviour. The current research investigates how various governing parameters affect heat transport within a v-shaped corrugated bottom triangular chamber consisting of Al2O3-H2O nanofluid. The vertical wall is kept at a uniform low temperature, the inclined wall is adiabatic, and the v-shaped corrugated bottom wall is exposed to constant high temperature. The Galerkin weighted residual approach, based on the method of finite elements, is used to solve the dimensionless PDE problems with boundary constraints. The isotherms, streamlines, and average Nusselt numbers are displayed for various Hartmann numbers (0 ≤ Ha ≤ 100), Rayleigh numbers (103 ≤ Ra ≤ 106), and nanoparticles volume fractions (0.01 ≤ ϕ ≤ 0.05). It demonstrates how the average Nu is affected by Ha, Ra, and ϕ. Furthermore, it has been discovered that the average Nu increases with Ra and ϕ but decreases with Ha. However, the Levenberg-Marquardt back propagation method is employed for the artificial neural network (ANN) approach on 120 datasets to estimate the average Nu, demonstrating an excellent correlation with simulated data and an average R2 of 0.99689. MATLAB software is used to calculate the ANN predictions.

PERAMALAN INDEKS HARGA SAHAM GABUNGAN (IHSG) MENGGUNAKAN METODE BACKPROPAGATION DAN RECURRENT NEURAL NETWORK (RNN)

Forecasting is an objective calculation and uses past data to predict something in the future. The economic activity of buying and selling shares is a common activity in this digital era. Investment in the capital market is currently popular among the general public. Forecasting in the digitalization era has undergone many developments, as well as its application not only in agriculture, but also in economics, politics and other fields. The development of forecasting in the economic field is growing rapidly with the use of more complex mathematical models and the application of information technology. Forecasting is an objective calculation and uses past data to predict something in the future. Utilization of artificial neural network (ANN) is one of the solutions for forecasting with more accurate results. One of the algorithms in ANN is the backpropagation algorithm. Apart from using other backpropagation forecasting methods that can be used is the deep learning (DL) method with a recurrent neural network algorithm. The backpropagation method produces an optimal MAPE of 13.28%. with an architectural model of seven input layer neurons, seven hidden layer neurons, and using the sigmoid activation function. The recurrent neural network method produces an optimal MAPE of 1.12%. with an architectural model of seven input layer neurons, ten hidden layer neurons, and using the tanh activation function. If the two optimal MAPEs of each method are compared, the Recurrent Neural Network Method produces more optimal forecasting compared to the backpropagation method.

Agent-Based Model for Situational Awareness in the Workplace: Enhancing Neural Networks with Direct Feedback Alignment

The integration of direct feedback alignment (DFA) into neural networks presents a paradigm shift in computational models, enhancing situational awareness within professional settings. This paper explores the application of DFA in agent-based computational models, demonstrating its efficiency and biological plausibility over traditional backpropagation methods. The research highlights the significant impact of direct feedback alignment on reinforcing situational awareness, evidenced through comprehensive simulations that show marked improvements in agents' ability to effectively navigate and comprehend complex work environments. The study suggests that direct feedback alignment revolutionizes neural network learning processes and significantly enhances cognitive aspects critical to employee safety and operational efficiency. It lays the groundwork for future investigations into the integration of neural computation techniques with organizational psychology and behavior, offering a new perspective on fostering safer, more aware, and efficient workplace environments. The potential application of direct feedback alignment across various professional scenarios opens new avenues for research in computational neuroscience, cognitive psychology, and organizational behavior, with a focus on optimizing human-environment interactions in complex systems.

Enhancing performance of shipboard photovoltaic grid-connected inverter through CRNN-LM-BP control optimized by particle swarm optimization of LCL parameters

The global renewable energy portfolio is experiencing a growing proportion of shipboard photovoltaic (PV) energy. Grid-connected system has become the development trend of photovoltaic power generation technology in marine applications because of its high energy utilization efficiency. However, in the grid-connected process, harmonic pollution is easily caused by time-varying marine environment disturbance, which affects the quality of grid-connected power. The current recurrent neural network Levenberg-Marquardt backpropagation (CRNN-LM-BP) method, a control methodology, is introduced in this paper to ensure the robustness, stability, and consistency of the system. In addition, the inductance-capacitance-inductance (LCL) filter is optimized using particle swarm optimization (PSO) to minimize the presence of high frequency harmonics. The system that has been developed possesses the ability to improve the accuracy of steady-state control, operate with high efficiency and effectiveness, and maintain a total harmonic distortion of 1.73 %. Concurrently, streamlining the Jacobian matrix in the LM algorithm expedites convergence. It enables the system to improve its dynamic response, minimize excess, and quickly return to a stable condition when faced with load variations, a wide range of filter parameter adjustments, and voltage fluctuations. The suggested method’s effectiveness and applicability are evaluated by comparing it to the traditional proportional-resonant control, CRNN-BP control, using Matlab/Simulink and the real-time simulator OPAL-RT5700..

Research on ground mobile robot trajectory tracking control based on MPC and ANFIS

This study focuses on the control strategy for a ground mobile robot (GMR) with independent three-axis six-wheel drive and four-wheel independent steering, performing double lane change trajectory tracking in complex scenarios. Initially, a dynamic model of the six-wheel independent drive and steering GMR was constructed. Utilizing Model Predictive Control (MPC) technology, the challenge of trajectory tracking at low speeds was effectively addressed. For high-speed conditions, by thoroughly analyzing the impact of the predictive time-domain, this study innovatively introduced an Adaptive Neuro-Fuzzy Inference System (ANFIS) to dynamically adjust the prediction horizon of the MPC. A novel trajectory tracking algorithm integrating MPC and ANFIS was developed, with the network structure being trained using backpropagation (BP) method and the least squares method. Compared to traditional MPC, this hybrid strategy significantly improves trajectory tracking accuracy and stability at high speeds, with computational efficiency increased by 48.65%. Additionally, the algorithm demonstrated excellent adaptability and control effectiveness in various rigorous tests, including different speed levels, complex steering paths, load changes, sudden obstacles, and variable terrain. A 70 km/h trajectory tracking experiment on a physical vehicle yielded a root mean square (RMS) error of 0.1904 m, verifying its superior tracking performance and practical reliability. This provides a pioneering solution for high-performance trajectory control of ground mobile robots.

Application of min-max normalization in backpropagation to detect early malnutrition

Malnutrition is the nutritional condition of toddlers which is characterized by a very thin condition, along with or without edema on the backs of both legs, weight according to body length or weight compared to height less than -3 standard deviations and/or upper arm circumference less than 11.5 cm in children aged 0-59 months. The contribution of this study is to apply min-max normalization to detect malnutrition using the Backpropagation method. This study uses the Backpropagation method with CRISP-DM, the input value of the artificial neural network is in an infinite range to a finite output value, which is in a range of 0 to 1. In order for this value to be met, a min-max normalization is carried out on the input value. The test results on this prototype achieved a maximum accuracy value of k= 1 with an accuracy value of 69%, Precision of 94% and Sensitivity of 69%. The results of the recommendation of the condition of children who are detected with malnutrition are very useful as an initial preventive measure in order to improve the nutritional condition of children with the right special treatment early and comprehensively.

A novel green learning artificial intelligence model for regional electrical load prediction

Load prediction is crucial to unit commitment and scheduling in electricity systems; however, load prediction models consume considerable electricity. Most conventional prediction methods are based on deep learning (DL) models, such as long short-term memory and gated recurrent units. The nonlinear activation functions and backpropagation mechanisms used in these methods result in high computational complexity and correspondingly high energy consumption. This paper proposes a prediction model based on a green learning (GL) framework aimed at reducing energy consumption. The proposed GL model replaces the activation functions conventionally used for feature extraction with a hybrid scheme combining categorical and numerical features, such as seasonal climate features and multi-autoregression terms. The proposed GL model also eliminates the backpropagation methods used to optimize hyperparameters by employing seasonal centroids and the quantile auto-regression forest algorithm for classification/regression. In a case study, the proposed GL model significantly outperformed conventional DL models in terms of energy consumption without compromising accuracy. The energy savings of the proposed GL model are equivalent to the annual electricity consumption of 211 ∼ 565 households, increasing with model penetration. This paper demonstrates the importance of energy-saving models in mitigating the linkage effect of industrial carbon emissions. It also explores policy implications for the implementation of GL models.

Pendekatan Neural Network dalam Peramalan Jumlah Penduduk Kota Semarang dengan Menggunakan Metode Backpropagation

The city of Semarang is one of the metropolitan cities that has a fairly dense population. During the years 2010-2021 the City of Semarang has a population fluctuation. It is necessary to make predictions of population data in order to plan the development in the City of Semarang can be better planned and can regulate the fluctuation of population in the future. In this study, the results of the prediction of the population of the City of Semarang were analyzed using the Neural Network approach with the backpropagation method. After training and testing, the best architectural model was obtained with 2 neurons on the input layer, 2 neurones on the hidden layer and 1 neuron on the output layer. Based on the results of the best architectural model, the MSE score was 9.39749 x 10-6 and the average MAPE value was 0.884552461%. The evaluation result with the MAPE value is very accurate because it is < 10%. In this study, the results of the forecast of the population of the city of Semarang in 2022-2025 consecutive are 1,863.121 people, 1,878.634 people, 1.891.865 people, and 1,902.947 people.

Hardware implementation of backpropagation using progressive gradient descent for in situ training of multilayer neural networks.

Neural network training can be slow and energy-expensive due to the frequent transfer of weight data between digital memory and processing units. Neuromorphic systems can accelerate neural networks by performing multiply-accumulate operations in parallel using nonvolatile analog memory. However, executing the widely used backpropagation training algorithm in multilayer neural networks requires information-and therefore storage-of the partial derivatives of the weight values preventing suitable and scalable implementation in hardware. Here, we propose a hardware implementation of the backpropagation algorithm that progressively updates each layer using in situ stochastic gradient descent, avoiding this storage requirement. We experimentally demonstrate the in situ error calculation and the proposed progressive backpropagation method in a multilayer hardware-implemented neural network. We confirm identical learning characteristics and classification performance compared to conventional backpropagation in software. We show that our approach can be scaled to large and deep neural networks, enabling highly efficient training of advanced artificial intelligence computing systems.

StocIPNet: A novel probabilistic interpretable network with affine-embedded reparameterization layer for high-dimensional stochastic inverse problems

The stochastic inverse problem (StocIP), which aims to align push-forward and observed output distributions by estimating probability distributions of unknown system inputs, often faces optimization challenges and the curse of dimensionality. A novel deep network called StocIPNet which comprises an affine-embedded reparameterization subnetwork (ReparNet) and a complex system metamodeling subnetwork (MetaNet) is proposed to alleviate these issues. The ReparNet subnetwork embeds the affine transformation to convert the statistical parameters of the physical random vector into learnable weights and biases, effectively implementing the reparameterization trick by separating random and deterministic elements in the stochastic sampling operation to preserve differentiability. In parallel, the MetaNet subnetwork offers a computationally efficient alternative to time-consuming forward solvers, facilitating the generation of push-forward distributions. The entire StocIPNet utilizes the kernel maximum mean discrepancy (MMD) as a distribution-free loss function, quantifying the discrepancy between push-forward and observed output distributions. By leveraging the ReparNet’s advantage of reformulating the sampling process as a differentiable transformation and combining two subnetworks seamlessly, the StocIP is reconfigured into a pure network training paradigm preserving differentiability perfectly, which allows for direct modeling and efficient inference of uncertainty within the network using automatic differentiation, backpropagation and gradient-based optimization methods, enabling ease of scaling to high-dimensional problems. The proposed framework has been theoretically demonstrated to be equivalent to the maximum likelihood method, ensuring its solid probabilistic interpretable foundation. The proposed framework is applied to perform stochastic model updating on a numerical and an experimental structure, which effectively demonstrates the framework’s remarkable effectiveness and high efficiency in treating high-dimensional StocIPs.

An artificial neural network and SCS–CN-based model for runoff estimation: a case study of the Peddavagu watershed

ABSTRACT The accurate prediction of runoff from rainfall events is crucial for effective water resource management, especially in regions with diverse climatic patterns like India. This study proposes a novel approach by integrating the soil conservation service (SCS)–curve number (CN) method with artificial neural networks (ANNs) to model rainfall–runoff relationships. In this research, an SCS–CN method is utilized to estimate initial runoff volumes, accounting for local soil and land-cover characteristics. Subsequently, an ANN-based model is developed to capture the complex nonlinear relationships between rainfall inputs and runoff outputs. Taking the rainfall of the watershed as the inputs, an ANN model was developed in MATLAB for runoff simulation at the main outlet of the Peddavagu watershed. A feed-forward back-propagation method is employed in the ANN model. The architecture with a 1-10-1 configuration based on tan–sig transfer function performed well in terms of MSE 0.019. The model efficiency was satisfactory with the coefficient of correlation (R), 0.96 for training, 0.98 for validation, and 0.98 for testing period. The overall value of R (0.96) indicates the utility of this ANN–SCS-based coupled model for rainfall–runoff simulation.

Azadirachta indica seed oil epoxidation process using carbonized melon seed peel catalyst; genetic algorithm coupled artificial neural network approach

The study used ANN-GA and RSM to predict the best process parameters for generating epoxide from Azadirachta indica seed oil (AISO). This procedure used carbonized sulphonated melon seed peel catalyst. FTIR, SEM, XRD, BET, and XRF measurements confirm the -SO3H group's attachment to the solid catalyst. The dependant variable was relative conversion to oxirane (RCO), while the independent parameters were catalyst dosage (0.6, 1.2, 1.8 wt %), time (4, 6, 8 h), and temperature (50°C, 60°C, 70°C). The ANN was evaluated using 11 backpropagation (BP) methods. Each method was examined with three input layer neurons for catalyst dosage, duration, and temperature. Ten neurons were in the hidden layer and one was in the output layer signifying RCO. The AISO epoxidation process forecast was most accurate using Bayesian regularisation. Simulated RSM and ANN models were built using experimental and algorithmic designs. The 3D plots showed that process parameters significantly affected RCO. R2 and MSE were used to evaluate model performance. For process forecasting, the ANN model (R2=0.9999, MSE=2.3404E-13) outperforms the RSM model (R2=0.9979, MSE=0.4688). Under the best RSM circumstances, RCO yield was 78.03 %. Additionally, the ANN and ANN-GA yielded 85.84 % and 92.51 %, respectively at optimal conditions of 0.6 wt % catalyst, 50°C temperature, and 6 h reaction time. However, all techniques optimized AISO and matched experimental results (RCO-77.41 %). FT-IR and GCMS characterizations of epoxy AISO corroborated the oxirane ring's attachment. The results show that ANN-GA is a reliable method for modelling and optimizing AISO epoxide production utilizing CSMSPC, encouraging sustainable development.

INTELLIGENT NUMERICAL METHOD FOR STUDYING MAXWELL WILLIAMSON NANOFLUID FLOW WITH ACTIVATION ENERGY

The use of artificial intelligence and its techniques has become increasingly widespread in recent times. It is being used to solve stiff non-linear equations. Additionally, nanofluids play a pivotal role in studying heat transfer. All of this was the motivation for doing this work. This work investigates a two-dimensional magnetohydrodynamic stretched flow (2D-MHDSF) of Maxwell Williamson nanofluid (MWNF) affected by bioconvection and activation energy numerically through Levenberg-Marquardt backpropagation method (LMBM)-based artificial neural network approach. The mathematical formulation for the problem was obtained through non-linear partial differential equations (PDEs). The leading PDEs were transmitted into non-linear ordinary differential equations by similarity transformation variables. The reference results for the 2D-MHDSF-MWNF model are produced by the Lobatto IIIA method through different scenarios of specific parameters for the flow velocity, fluid temperature, nanoparticle concentration, and motile density profiles. Using obtained results as a dataset to apply the testing, training, and validation steps of the suggested LMBM for the 2D-MHDSF-MWNF model. The mean squared error, analysis of regression, and error histograms are presented to prove the efficiency and precision of the proposed method. The numerical results of LMBM are displayed as a study of the effects of different physical factors on flow dynamics for 2D-MHDSF-MWNF.