Back-propagation Artificial Neural Network Research Articles

Artificial intelligence and machine learning in optics: tutorial

Across the spectrum of scientific inquiry and practical applications, the emergence of artificial intelligence (AI) and machine learning (ML) has comprehensively revolutionized problem-solving methodologies. This tutorial explores key aspects of AI/ML and their remarkable role in augmenting the capabilities of optics and photonics technologies. Beginning with fundamental definitions and paradigms, the tutorial progresses to classical machine learning algorithms, with examples employing support vector machines and random forests. Extensive discussion of deep learning encompasses the backpropagation algorithm and artificial neural networks, with examples demonstrating the applications of dense and convolutional neural networks. Data augmentation and transfer learning are examined next as effective strategies for handling scenarios with limited datasets. Finally, the necessity of alleviating the burden of data collection and labeling is discussed, motivating the investigation of unsupervised and semi-supervised learning strategies as well as the utilization of reinforcement learning. By providing a structured exploration of AI/ML techniques, this tutorial equips researchers with the essential tools to begin leveraging AI’s transformative potential within the expansive realm of optics and photonics.

Support Vector Machine, Naive Bayes, and Artificial Neural Network Back Propagation Comparison in Detecting Brain Tumor

Support Vector Machine, Naive Bayes, and Artificial Neural Network Back Propagation Comparison in Detecting Brain Tumor

Predicting the geospatial distribution of Chinese rice nutrient element in regional scale for the geographical origin—A case study on the traceability of Japonica rice

AbstractEffective geographical origin discrimination of Chinese rice requires a large database of samples to ensure sufficient data for origin verification at a regional scale. In this study, environmental similarity was used to establish a spatial database of rice nutrient element, and then the validity of the database was verified using the back propagation artificial neural networks modeling (BPNN). The spatial distribution model of 14 rice nutrient element (Al, Ba, Ca, Cu, Cr, Fe, K, Mg, Mn, Mo, Na, Ni, Rb, and Zn) on regional scale was built using an environmental similarity method for the first time. Elemental concentrations of 692 samples were used to build a simulated geographical origin prediction model for northeastern (N‐E), middle to lower Yangtze River plain (Y‐R), southwestern (S‐W), and southeastern (S‐E) in China. The results indicated that the performance of the environmental similarity model for these four growing regions was S‐W > N‐E > S‐E > Y‐R based on the lowest ranking root mean square error (RMSE) for each region. For example, the RMSEs of Zn in S‐W, N‐E, S‐E, and Y‐R regions were 2.0, 2.4, 2.7, and 3.7 mg/kg, respectively. A case study on the traceability of Japonica rice was shown that Japonica rice could be discriminated with higher origin accuracy using a simulated database (91.8%) than by the actual database (87.0%) using the BPNN model. This indicates that a simulated rice element database could improve the accuracy of geographical origin discrimination for Chinese rice and potentially be applied to other large national‐scale crop datasets.

Classification of Malignancy of Lung Cancer Using Backpropagation Algorithm on CT-Scan Images

In this study, we investigate the classification of lung cancer CT scan images based on malignancy level using a backpropagation artificial neural network (ANN). Lung cancer is a deadly disease characterized by the growth of abnormal lung cells. The proposed method involves preprocessing to enhance image quality, followed by feature extraction using the Gray Level Co-occurrence Matrix (GLCM) method with angle variations of 0°, 45°, 90°, 135°, and d=1. The extracted features include energy, contrast, correlation, and homogeneity. The energy value range in malignant cancer is 0.27 to 0.81, while in benign cancer it is 0.26 to 0.73. The contrast in benign cancer ranges from 1.38 to 11.87, while in malignant cancer it is 1.47 to 13.67. The image correlation for malignant cancer is between 0.63 to 0.94, while for benign cancer it is 0.69 to 0.96. Homogeneity in malignant cancer has a value range between 0.67 to 0.91, while in benign cancer it ranges from 0.70 to 0.92. The classification of lung cancer malignancy is restricted to benign and malignant levels using a network architecture of [4 10 2], maximum iteration of 100000, and learning rate of 0.001. The accuracy of the testing data from the ANN is between 90% and 100%. These results demonstrate the effectiveness of the GLCM method and backpropagation algorithm in accurately classifying the malignancy level of lung cancer, which could aid in the early detection and treatment of the disease.

Application of artificial intelligence combined with near infrared spectroscopy in the continuous counter-current extraction process of Angelica dahurica formula granules

The establishment of near infrared (NIR) spectroscopy model mostly relies on chemometrics, and spectral analysis combined with artificial intelligence (AI) provides a new way of thinking for pharmaceutical quality inspection, new algorithms such as back propagation artificial neural networks (BP-ANN) and swarm intelligence optimization algorithms such as sparrow search algorithm (SSA) provide core technical support. In order to explore the application of AI in the pharmaceutical field, in this study, Angelica dahurica formula granules with a relatively complex system were selected as the research object. Quantitative analysis models were established by using partial least squares regression (PLSR) with a micro-NIR spectrometer, and BP-ANN modeling results were compared. For the best PLSR models of six characteristic components in the continuous counter-current extract of Angelica dahurica, R2v of imperatorin was lower than 0.90, and the RPD values of imperatorin, phellopterin, and isoimperatorin were even lower than 1. When the prediction model established by SSA-BP-ANN was used for quantitative analysis, R2v of six components were all higher than 0.92, and the RPD values all higher than 1.5, which proved that the BP-ANN method was better than PLSR. This study confirmed that in the continuous counter-current extraction progress of Angelica dahurica formula granules, the use of micro-NIR spectrometer combined with AI could realize the rapid prediction of the contents of six characteristic components. The comparison results provided a scientific reference for the process analysis and on-line monitoring in the production process of traditional Chinese medicine by micro-NIR spectrometer combined with AI.

Collaboration of bacterial consortia for biodegradation of high concentration phenol and potential application of machine learning

Mixed culture of microorganisms is an effective method to remove high concentration of phenol in wastewater. At present, it is still a challenge for microorganisms to remove high-concentration phenol from wastewater. In this study, a phenol-degrading consortium was isolated, which could rapidly degrade 1800 mg/L phenol within 30 h, and the highest phenol degradation concentration was 2000 mg/L. Further exploration of how microbial consortium cooperates to promote phenol biodegradation was studied: the core bacteria of the microbial consortium was relatively stable during phenol degradation; the bacteria could improve the adaptability to environment and metabolic ability of phenol, by producing more surfactants and betaine, thereby improving the degradation rate. The determination coefficient (R2) in the machine learning model showed that the back propagation artificial neural network (BP-ANN) can predict the biodegradation of phenol under different conditions, saving time and economic costs. This study explains how microbial consortium cooperates to degrade phenol from the aspects of microbial consortium composition and metabolic analysis, which provides a theoretical basis for mixed culture microorganisms to degrade pollutants.

Improving support vector machine and backpropagation performance for diabetes mellitus classification

Diabetes mellitus (DM) is a glucose disorder disease in the human body that contributes significantly to the high mortality rate. Various studies on early detection and classification have been conducted as a DM prevention effort by applying a machine learning model. The problems that may occur are weak model performance and misclassification caused by imbalanced data. The existence of dominating (majority) data causes poor model performance in identifying minority data. This paper proposed handling the problem of imbalanced data by performing the synthetic minority oversampling technique (SMOTE) and observing its effect on the classification performance of the support vector machine (SVM) and Backpropagation artificial neural network (ANN) methods. The experiment showed that the SVM method and imbalanced data achieved 94.31% accuracy, and the Backpropagation ANN achieved 91.56% accuracy. At the same time, the SVM method and balanced data produced an accuracy of 98.85%, while the Backpropagation ANN method and balanced data produced an accuracy of 94.90%. The results show that oversampling techniques can improve the performance of the classification model for each data class.

Improving support vector machine and backpropagation performance for diabetes mellitus classification

Diabetes mellitus is a glucose disorder disease in the human body that contributes significantly to the high mortality rate. Various studies on early detection and classification have been conducted as a diabetes mellitus prevention effort by applying a machine learning model. The problems that may occur are weak model performance and misclassification caused by imbalanced data. The existence of dominating (majority) data causes poor model performance in identifying minority data. This paper proposed handling the problem of imbalanced data by performing the synthetic minority oversampling technique (SMOTE) and observing its effect on the classification performance of the support vector machine (SVM) and Backpropagation artificial neural network (ANN) methods. The experiment showed that the SVM method and imbalanced data achieved 94.31% accuracy, and the Backpropagation ANN achieved 91.56% accuracy. At the same time, the SVM method and balanced data produced an accuracy of 98.85%, while the Backpropagation ANN method and balanced data produced an accuracy of 94.90%. The results show that oversampling techniques can improve the performance of the classification model for each data class.

An aeration requirements calculating method based on BOD5 soft measurement model using deep learning and improved coati optimization algorithm

Conventionally, the supply-side is over-aerated to meet the oxygen demand of biological wastewater treatment in wastewater treatment plants (WWTPs). Such systems do not consider the various influent loadings and actual oxygen consumption. It cannot respond effectively to fluctuating influent, resulting in unnecessary aeration and energy wastage. In this study, we developed a novel approach to calculate aeration requirements in real time. Firstly, the oxygen demand formula for biological nitrogen and phosphorus removal is modified by considering realistic factors (temperature, oxygen partial pressure and so on) to construct the aeration demand formula. Then, to address the problem that BOD5 in aeration formula is difficult to measure in real time, a novel hybrid model integrating Fuch mapping, reverse learning strategy, Coati Optimization Algorithm (COA), and Backpropagation Artificial Neural Network (BP-ANN) is proposed in this study. Finally, aeration requirements are calculated based on real-time water quality data. Experimental results show that the ICOA-BP model with 9 neurons in hidden layer predicted the optimum BOD5 concentration by achieving a 25.146 mean absolute error and a 0.735 coefficient of determination (all-R2) in all datasets. The optimization led to a 15.164 % aeration savings and 38.942 % energy savings. The proposed method of aeration calculation is automated and intelligent, responding to fluctuating changes in influent quality and reducing unnecessary aeration and energy consumption while meeting the requirements of effluent quality.

Electronic Nose Based on Sensor Array for Classification of Beef and Rat Meat Using Backpropagation Artificial Neural Network Method

Electronic Nose Based on Sensor Array for Classification of Beef and Rat Meat Using Backpropagation Artificial Neural Network Method

A Railway Roadbed Deformation Monitoring System Using Deep Learning and AI Intelligent Technology

The author has designed an automated monitoring system for settlement and deformation of high-speed railway subgrade. Firstly, an automated monitoring system is designed based on sensors, data collection and transmission, client tracking and querying, monitoring result processing, automated warning, manual monitoring data analysis, monitoring data analysis and evaluation. Secondly, the system software and hardware are designed, the author calculated and analyzed engineering examples from the perspective of practical applications using artificial neural network methods, obtained corresponding deformation analysis models, and predicted deformation. Finally, the practical application of the system was analyzed for its effectiveness. The experimental results indicate that, the BP artificial neural network method is used to model and predict the Deformation monitoring data. On the premise of 20 learning samples and 4 prediction samples, the Root-mean-square deviation of the prediction is 0.32mm, which shows that the deformation prediction using BP model is feasible and effective in a certain precision range. It has been proven that the application of artificial neural network methods in monitoring and prediction of practical engineering has certain practical significance.

Prediction of COD in industrial wastewater treatment plant using an artificial neural network

In this investigation, the modeling of the Aksaray industrial wastewater treatment plant was performed using artificial neural networks with various architectures in the MATLAB software. The dataset utilized in this study was collected from the Aksaray wastewater treatment plant over a 9-month period through daily records. The treatment efficiency of the plants was assessed based on the output values of chemical oxygen demand (COD) output. Principal component analysis (PCA) was applied to furnish input for the Feedforward Backpropagation Artificial Neural Networks (FFBANN). The model’s performance was evaluated using the Mean Squared Error (MSE), the Mean Absolute Error (MAE) and correlation coefficient (R2) parameters. The optimal architecture for the neural network model was determined through several trial and error iterations. According to the modeling results, the ANN exhibited a high predictive capability for plant performance, with an R2 reaching up to 0.9997 when comparing the observed and predicted output variables.

Estimating Shear Strength of Marine Soft Clay Sediment: Experimental Research and Hybrid Ensemble Artificial Intelligence Modeling

In the design of offshore engineering foundations, a critical consideration involves determining the peak shear strength of marine soft clay sediment. To enhance the accuracy of estimating this value, a database containing 729 direct shear tests on marine soft clay sediment was established. Employing a machine learning approach, the Particle Swarm Optimization algorithm (PSO) was integrated with the Adaptive Boosting Algorithm (ADA) and Back Propagation Artificial Neural Network (BPANN). This novel methodology represents the initial effort to employ such a model for predicting the peak shear strength of the soil. To validate the proposed approach, four conventional machine learning algorithms were also developed as references, including PSO-optimized BPANN, Support Vector Machine (SVM), BPANN, and ADA-BPANN. The study results show that the PSO-BPANN model, which has undergone optimization via Particle Swarm Optimization (PSO), has prediction accuracy and efficiency in determining the peak shear performance of marine soft clay sediments that surpass that offered by traditional machine learning models. Additionally, a sensitivity analysis conducted with this innovative model highlights the notable impact of factors such as normal stress, initial soil density, the number of drying–wetting cycles, and average soil particle size on the peak shear strength of this type of sediment, while the impact of initial soil moisture content and temperature is comparatively minor. Finally, an analytical formula derived from the novel algorithm allows for precise estimation of the peak shear strength of marine soft clay sediment, catering to individuals lacking a background in machine learning.

Implementation of Backpropagation Artificial Neural Network for Electricity Load Forecasting in Jember District

The increase in population and various kinds of human activities in the world has made it possible for changes to increase the need for electrical power with demand that is not the same at any time. Based on this description, this research will propose research on the theme of electricity load forecasting as a preventive measure to determine future electricity load needs. Research was assisted using MATLAB data processing software to process research data. Three forecasting models were carried out, namely day, night and day-night conditions. From these three forecasting models, parameters such as epoch, number of input layers, number of hidden layers, activation function, and etc. The data is divided into two parts, training data and test data with a ratio of 70: 30. Test results using the backpropagation artificial neural network method show the highest MSE values for the three forecasting models, day, night, and day-night, are, 0.0039, 0.0041, and 0.002 while the lowest MSE values were in the three models are, 6.77E-04, 0.001, and 0.0011.

Study on hot deformation behavior and recrystallization mechanism of an Al-6.3Zn-2.5Mg-2.6Cu-0.11Zr alloy based on machine learning

The data from thermal simulation plays a significant role in understanding the deformation behavior and deformation mechanisms of materials, and provides guidance for the design of hot deformation processes. In this paper, thermal simulation experiments were performed on an Al-6.3Zn-2.5Mg-2.6Cu-0.11Zr alloy, and the microstructure after hot deformation were characterized by Electron Backscatter Diffraction (EBSD) technique. The thermal simulation data and microstructural characteristics were analyzed and studied using machine learning technique: the Backpropagation Artificial Neural Network (BP-ANN) method, correlation analysis and K-means clustering analysis. A model for predicting not only flow stress but also dislocation density was established using the BP-ANN method, which exhibits good prediction accuracy. The R² and AARE values are 90.19 % and 6.87 % for the prediction of dislocation density, and are 98.67 % and 6.96 % for the prediction of flow stress. Subsequently, correlation analysis was performed to investigate the relationship between energy dissipation efficiency η and deformation parameters as well as microstructural characteristics, revealing that energy dissipation rate is correlated positively with the content of dynamical recrystallization (DRX) and high angle grain boundary (coefficients of 0.90 and 0.82) and negatively with the dislocation density and the Zener-Hollomon parameter (coefficients of −0.98 and −0.97). Additionally, K-means clustering analysis was applied to study the relationship between the energy dissipation efficiency and softening mechanisms, it was found that the softening mechanisms changed with the variation of η value, dynamic recovery (DV), geometric DRX (GDRX) and continuous DRX (CDRX) dominated when η>0.38; DV, CDRX and discontinuous DRX(DDRX) dominated when 0.28<η<0.38; DV dominated when η<0.28. The results show that machine learning techniques are useful tools in mining thermal simulation data, and more information can be obtained than traditional data mining methods.

Application of Predictions for the Total Productivity of Competency Certificates Electrical Engineering Staff at PT. Eleska Iatki Using Artificial Neural Network Backpropagation Algorithm

Application of Predictions for the Total Productivity of Competency Certificates Electrical Engineering Staff at PT. Eleska Iatki Using Artificial Neural Network Backpropagation Algorithm

Rapid quantification of single component oil in perilla oil blends by ultraviolet–visible spectroscopy combined with chemometrics

As a unconventional oil, perilla oil is much more expensive than conventional oils since it has the highest content of α-linolenic acid among vegetable oils. Thus the adulteration of perilla oil is serious, which needs to be solved. In this study, the single component oil in perilla oil blends were first quantitatively analyzed by ultraviolet–visible (UV–vis) spectroscopy combined with chemometric methods. Soybean oil and palm oil were added into perilla oil to form binary and ternary perilla oil blends. Partial least squares (PLS), back propagation-artificial neural network (BP-ANN), support vector regression (SVR) and extreme learning machine (ELM) were compared and the best model was selected for calibration. In order to improve the prediction performance of the calibration model, ten preprocessing methods and five variable selection methods were investigated. Results show that PLS was the best calibration method for binary and ternary perilla oil blends. For binary perilla oil blends, the correlation coefficients of prediction (Rp) obtained by PLS were both above 0.99, which does not need preprocessing and variable selection. For ternary perilla oil blends, after the best continuous wavelet transform (CWT) preprocessing and discretized whale optimization algorithm (WOA) variable selection, the Rp values obtained by the best model CWT-WOA-PLS were all above 0.97. This research provides a common framework for calibration of perilla oil blends, which maybe a promising method for quality control of perilla oil in industry.

APPLICATION OF MACHINE LEARNING MODELS AND GSA METHOD FOR DESIGNING STUD CONNECTORS

The design of stud connectors is aided by determining the relationship between shear strength and the input variables (number, diameter, height, tensile strength and elastic modulus of the studs, and compressive strength and elastic modulus of the concrete) that influence strength. Since strength is nonlinearly related to the influencing variables, which makes the predictions of the relevant empirical equations unreliable, the use of machine learning (ML) models is preferred. The prediction results of eight machine learning models were evaluated, including linear regression (LR1), ridge regression (RR), lasso regression (LR2), back-propagation artificial neural network (BP ANN), genetic algorithm optimized BP ANN (GA-BP ANN), extreme learning machines (ELM), random forests (RF), and support vector machines (SVM). The results show that the GA-BP ANN model is the most accurate model for prediction with a mean absolute percentage error (MAPE) of 6.17% and an R2 of 0.9599. Based on the GA-BP ANN model and the global sensitivity analysis (GSA) method, a new parameter importance analysis method was developed to compare the magnitude of the effect of different input variables on strength. It was found that stud diameter had the greatest effect on shear strength.

Operational Reliability Analysis of Turbine Blisk Using an Enhanced Moving Neural Network Framework

As one of the key components of an aeroengine, turbine blisk endures complex coupling loads under a harsh operational environment so that the reliability of turbine blisk directly influences the safe operation of aeroengine. It is urgent to precisely perform the reliability estimation of a complex blisk structure. To address this issue, an enhanced Moving Neural Network Framework (MNNF) is proposed by integrating compact support region theory, improve sooty tern optimization algorithm (ISTOA), and Bayesian regularization strategy into artificial neural network. The compact support region theory is applied to select the efficient samples for modeling from the training samples set, the ISTOA is to determine the optimal compact support region, and Bayesian regularization thought is utilized to improve the generalization ability of neural network model. The operational reliability assessment of aeroengine blisk is performed with the consideration of transient loads to verify the proposed MNNF method. It is shown that the reliability degree of turbine blisk stain is 0.9984 when the allowable value is 5.2862 × 10−3 m. In line with the comparison of methods, the developed MNNF approach has 0.99738 in root means square error, 3.1634 × 10−4 m in goodness of fit, 0.423 s in modeling time, 99.99% in simulation precision, and 0.496 s in simulation time under 10,000 simulations, which are superior to all other methods (i.e., 99.96%, 99.91%, 99.93%, 99.97%, and 99.97% in simulation precision and 16.27%, 4.82%, 30.07%, 39.87%, and 23.59% in simulation efficiency, for the response surface method (RSM), Kriging, support vector machine (SVM), back propagation-artificial neural network (BP-NN), and BP-NN based on particle swarm optimization (BP-PSO) methods, respectively). It is demonstrated that the MNNF method holds excellent modeling and simulation performances. The efforts of this study provide promising tools and insights into the reliability design of complex structures, and enrich and develop reliability theory.

Settlement Forecast of Marine Soft Soil Ground Improved with Prefabricated Vertical Drain-Assisted Staged Riprap Filling

By comparing different settlement forecast methods, eight methods were selected considering the creep of marine soft soils in this case study, including the Hyperbolic Method (HM), Exponential Curve Method (ECM), Pearl Growth Curve Modeling (PGCM), Gompertz Growth Curve Modeling (GGCM), Grey (1, 1) Model (GM), Grey Verhulst Model (GVM), Back Propagation of Artificial Neural Network (BPANN) with Levenberg–Marquardt Algorithm (BPLM), and BPANN with Gradient Descent of Momentum and Adaptive Learning Rate (BPGD). Taking Lingni Seawall soil ground improved with prefabricated vertical drain-assisted staged riprap filling as an example, forecasts of the short-term, medium-term, long-term, and final settlements at different locations of the soft ground were performed with the eight selected methods. The forecasting values were compared with each other and with the monitored data. When relative errors were between 0 and −1%, both the forecasting accuracy and engineering safety were appropriate and reliable. It was concluded that the appropriate forecast methods were different not only due to the time periods during the settlement process, but also the locations of soft ground. Among these methods, only BPGD was appropriate for all the time periods and locations, such as at the edge of the berm, and at the center of the berm and embankment.