Back Propagation Neural Network Method Research Articles

An Inflation Rate Prediction Based on Backpropagation Neural Network Algorithm

This article aims to predict the inflation rate in Samarinda, East Kalimantan by implementing an intelligent algorithm, Backpropagation Neural Network (BPNN). The inflation rate data was obtained from the Provincial Statistics Bureau of Samarinda https://samarindakota.bps.go.id/ for the period January 2012 to January 2017. The method used to measure accuracy algorithm prediction was the mean square error (MSE). Based on the experiment results, the BPNN method with architectural parameters of 5-5-5-1; the learning function was trainlm; the activation functions were logsig and purelin; the learning rate was 0.1 and able to produce a good level of prediction error with an MSE value of 0.00000424. The results showed that the BPNN algorithm can be used as an alternative method in predicting inflation rates in order to support sustainable economic growth, so that it can improve the welfare of the people in Samarinda, East Kalimantan.

Improvement of sap flow estimation by including phenological index and time-lag effect in back-propagation neural network models

Sap flow is an important indicator reflecting water transport in a soil-plant-atmosphere continuum system. Therefore, identifying the effects of different factors on sap flow is imperative for understanding its physiological responses to environmental conditions and other associated ecological processes. However, conventional statistical methods have produced unsatisfactory results due to the complex and nonlinear relationship between sap flux density (vs) and its driving factors. This study illustrated the utility of the back-propagation (BP) neural network method for sap flow estimation and compared the performance of BP models with the multiple-linear regression (MLR) technique. Based on the measured sap flow of Pinus massoniana in the field, three-layer BP models trained by the Levenberg-Marquardt algorithm were developed with an architecture of 4-10-1, corresponding to four, ten and one nodes in the input, hidden and output layers, respectively. The BP models were trained and validated in the MATLAB environment with four different combinations of air temperature (Ta), relative humidity (RH), average net radiation (ANR) and the phenological index. High degrees of correlation were observed between the measured and simulated results by BP, and the coefficients of determination (R2) and fitting accuracies (Acc) (greater than 0.9 and 80%, respectively) were higher than the corresponding values from the MLR (0.78 and 69%, respectively). Furthermore, the performance of the BP models could be greatly improved by including the phenological index and a time-lag effect, thereby suggesting that these two factors were crucial variables in modelling vs by BP approaches. The BP models were also tested with cross-validation method and 50% of the collected data that were not used in model development. We conclude that the BP model had a higher degree of accuracy in predicting sap flow due to its superior performance for complicated, nonlinear and uncertain processes, especially with inclusion of the phenological index and time-lag effect.

International roughness index prediction based on multigranularity fuzzy time series and particle swarm optimization

The effective prediction of pavement performance trends can help in achieving the cost-effective management of pavements over their service life. The international roughness index (IRI) is a widely used pavement performance index, which can be considered as a time-dependent variable in terms of scientific modeling. This research aims to develop an innovative IRI prediction model based on fuzzy-trend time-series forecasting and particle swarm optimization (PSO) techniques. Raw datasets extracted from the Long-Term Pavement Performance database are used for model training, testing, and performance assessment. First, IRI values are divided into different granular spaces, which are considered as the principal factor and subfactors. In addition, the multifactor interval division method is proposed according to the principle of the automatic clustering technique. Next, a second-order fuzzy-trend model and fuzzy-trend relationship classification method are proposed to predict the fuzzy-trend of each factor. Then, the fuzzy-trend states for multiple granular spaces are generated while giving full consideration to various uncertainties. Finally, the PSO technique is used to optimize the performance model while carrying out future IRI forecasting. Comparative experiments are performed using more than 20,000 data items from different regions to verify the effectiveness of the proposed method. The experimental results indicate that the proposed method outperforms other approaches including the polynomial fitting, autoregressive integrated moving average, and backpropagation neural network methods in terms of the root mean squared error (0.191) and relative error (6.37%).

Retrieving volume FeO and TiO2 abundances of lunar regolith with CE-2 CELMS data using BPNN method**A contributed paper of the International Symposium on Lunar and Planetary Science (ISLPS) on 2018 June 12–15 at Macau University of Science and Technology.

The volume FeO and TiO2 abundances (FTAs) of lunar regolith can be more important for understanding the geological evolution of the Moon compared to the optical and gamma-ray results. In this paper, the volume FTAs are retrieved with microwave sounder (CELMS) data from the Chang’E-2 satellite using the back propagation neural network (BPNN) method. Firstly, a three-layered BPNN network with five-dimensional input is constructed by taking nonlinearity into account. Then, the brightness temperature (TB) and surface slope are set as the inputs and the volume FTAs are set as the outputs of the BPNN network. Thereafter, the BPNN network is trained with the corresponding parameters collected from Apollo, Luna, and Surveyor missions. Finally, the volume FTAs are retrieved with the trained BPNN network using the four-channel TB derived from the CELMS data and the surface slope estimated from Lunar Orbiter Laser Altimeter (LOLA) data. The rationality of the retrieved FTAs is verified by comparing with the Clementine UV-VIS results and Lunar Prospector (LP) GRS results. The retrieved volume FTAs enable us to re-evaluate the geological features of the lunar surface. Several important results are as follows. Firstly, very-low-Ti (<1.5 wt.%) basalts are the most spatially abundant, and the surfaces with TiO2 > 5 wt.% constitute less than 10% of the maria. Also, two linear relationships occur between the FeO abundance (FA) and the TiO2 abundance before and after the threshold, 16 wt.% for FA. Secondly, a new perspective on mare volcanism is derived with the volume FTAs in several important mare basins, although this conclusion should be verified with more sources of data. Thirdly, FTAs in the lunar regolith change with depth to the uppermost surface, and the change is complex over the lunar surface. Finally, the distribution of volume FTAs hints that the highlands crust is probably homogeneous, at least in terms of the microwave thermophysical parameters.

Quality Characteristics Analysis and Remaining Shelf Life Prediction of Fresh Tibetan Tricholoma matsutake under Modified Atmosphere Packaging in Cold Chain.

Tricholoma matsutake (T. matsutake) growing in Tibet is very popular for its high economic and medicinal value, but fresh T. matsutake has an extremely short shelf life. The shelf life of T. matsutake is complex, influenced by product characteristics, surrounding environmental conditions, and spoilage development. The objective of this work was to study the quality characteristics of fresh T. matsutake during its shelf life period in modified atmosphere packaging (MAP) conditions and establish its remaining shelf life prediction models in a cold chain. In this study, we measured and analyzed quality indicators of fresh T. matsutake, including hardness (cap, stipe), color, odor of sensory characteristics, pH, soluble solids content (SSC), and moisture content (MC) of physical and chemical characteristics under the temperature condition of 4 °C and relative humidity (RH) of 90%. The sensory evaluation results showed that the odor indicator in sensory characteristics was more sensitive to the freshness of T. matsutake. The changes of pH, SSC, and MC were divided into three periods to analyze the physiological changes of T. matsutake. The cap spread process could affect the changes of pH, SSC, and MC in period S1, and they changed gradually in period S2. In the period S3, they changed complicatedly because of deterioration. The remaining shelf life prediction model of T. matsutake was established by the back propagation (BP) neural network method to quantify the relationship between the quality indicators and the remaining shelf life. The shelf life characteristics are complex, which were optimized by correlation analysis. Significant benefits of this work are anticipated on the transportation and preservation of fresh T. matsutake to the market and the reduction of its losses in the postharvest chain.

Credit Scoring Kelayakan Debitur Menggunakan Metode Hybrid ANN Backpropagation dan TOPSIS

Credit is one of the common practices that provide benefits for financial or non-financial institutions. However on the other hand, aid loans also have higher risks if the institutions give the wrong decision in giving a loan. Credit Scoring is one of techniques that can determine whether it is feasible to given a loan or not. The selection of a credit scoring model greatly determines the value in classifying credit that is feasible or not to giving a loan. Decision Support System (DSS) is one system that can be used to overcome this problem. The advantages of DSS are being able to overcome the problems that have semi-structured and unstructured data. In this study, DSS was supported by using Artificial Neural Network Backpropagation method and TOPSIS method to find the priority for seeking eligibility. Accuracy results obtained in this study reached 98,69% with the number of iteration is 300, the number of training data is 30, neuron hidden 12 and error tolerance is 0.001. TOPSIS method succeeded in ranking 185 data selected as recipients of credit.&#x0D; Keywords:Credit Scoring, Decision Support System (DSS), Artificial Neural Network (ANN), Backpropagation, TOPSIS.

Case-based reasoning model based on attribute weights optimized by genetic algorithm for predicting end temperature of molten steel in RH

Temperature control is the key of Ruhrstahl-Heraeus (RH) process in steelmaking plant. The accuracy of RH control model greatly affects the molten steel temperature fluctuation. To obtain RH control model with higher accuracy, an improved case-based reasoning (CBR) model based on attribute weights optimized by genetic algorithm (GA) was proposed. The fitness function in GA was determined according to the prediction accuracy of end temperature of molten steel in RH; then, GA is used to optimize all the attribute weights based on known case base. An improved CBR model that contains the optimized attribute weights was applied to predict end temperature of molten steel in RH, and the prediction accuracy was calculated. Four methods, CBR based on attribute weights optimized by GA (CBR–GA), ordinary CBR, back propagation neural network (BPNN) and multiple linear regression (MLR) method were employed for comparison. The results show that in the error range of [− 3 °C, 3 °C], [− 5 °C, 5 °C], [− 7 °C, 7 °C] and [− 10 °C, 10 °C], the prediction accuracy of CBR–GA was improved by 19.99%, 28.19%, 27.11% and 16.3%, respectively, than that of MLR. Compared with BPNN, the prediction accuracy increased by 3.22%, 7.44%, 5.29% and 2.40%, respectively. Compared with ordinary CBR, the accuracy increased by 5.43%, 5.80%, 4.66% and 2.27%, respectively.

Integrating Growth and Environmental Parameters to Discriminate Powdery Mildew and Aphid of Winter Wheat Using Bi-Temporal Landsat-8 Imagery

Monitoring and discriminating co-epidemic diseases and pests at regional scales are of practical importance in guiding differential treatment. A combination of vegetation and environmental parameters could improve the accuracy for discriminating crop diseases and pests. Different diseases and pests could cause similar stresses and symptoms during the same crop growth period, so combining growth period information can be useful for discerning different changes in crop diseases and pests. Additionally, problems associated with imbalanced data often have detrimental effects on the performance of image classification. In this study, we developed an approach for discriminating crop diseases and pests based on bi-temporal Landsat-8 satellite imagery integrating both crop growth and environmental parameters. As a case study, the approach was applied to data during a period of typical co-epidemic outbreak of winter wheat powdery mildew and aphids in the Shijiazhuang area of Hebei Province, China. Firstly, bi-temporal remotely sensed features characterizing growth indices and environmental factors were calculated based on two Landsat-8 images. The synthetic minority oversampling technique (SMOTE) algorithm was used to resample the imbalanced training data set before model construction. Then, a back propagation neural network (BPNN) based on a new training data set balanced by the SMOTE approach (SMOTE-BPNN) was developed to generate the regional wheat disease and pest distribution maps. The original training data set-based BPNN and support vector machine (SVM) methods were used for comparison and testing of the initial results. Our findings suggest that the proposed approach incorporating both growth and environmental parameters of different crop periods could distinguish wheat powdery mildew and aphids at the regional scale. The bi-temporal growth indices and environmental factors-based SMOTE-BPNN, BPNN, and SVM models all had an overall accuracy high than 80%. Meanwhile, the SMOTE-BPNN method had the highest G-means among the three methods. These results revealed that the combination of bi-temporal crop growth and environmental parameters is essential for improving the accuracy of the crop disease and pest discriminating models. The combination of SMOTE and BPNN could effectively improve the discrimination accuracy of the minor disease or pest.

Partus referral classification using backpropagation neural network

The number of maternal mortalities in Indonesia keeps increasing significantly. The same thing occurs to the number of fetal mortalities which is not less than maternal ones. It is a given knowledge that the condition of expectants is mostly vulnerable to various disorders. Therefore, there is a high possibility of the occurrence of any maternal and fetal mortality. Things most likely to go wrong when the partus occurs and if the disorders are not treated as fast as possible, it will be very dangerous for the expectants. To minimize such occurrences, medical officers have been using a certain method named screening to detect which partus case which potentially has high risk before it is too late. Nevertheless, the results of screening are not always accurate. In this paper, we are going to come up with any other way, which is data mining computation. Data mining has known to be able to help human solving problems as well as a promising region for data classification. Most medical problems have been classified using data mining. Backpropagation Neural Network method will be used in this paper to classify partus referral cases. Thus, Backpropagation Neural Network will be compared to other algorithms to know if it is a better algorithm for this case based on real data of cohort register. The result of the researches confirms that Backpropagation Neural Network is the best of all three methods (Backpropagation Neural Network, C4.5 and Naive Bayes Classifier) used to classify partus referral.

Hybrid backpropagation neural network-particle swarm optimization for seismic damage building prediction

The evaluation of the vulnerability of buildings to earthquakes is of prime importance to ensure a good plan can be generated for the disaster preparedness to civilians. Most of the attempts are directed in calculating the damage index of buildings to determine and predict the vulnerability to certain scales of earthquakes. Most of the solutions used are traditional methods which are time consuming and complex. Some of initiatives have proven that the artificial neural network methods have the potential in solving earthquakes prediction problems. However, these methods have limitations in terms of suffering from local optima, premature convergence and overfitting. To overcome this challenging issue, this paper introduces a new solution to the prediction on the seismic damage index of buildings with the application of hybrid back propagation neural network and particle swarm optimization (BPNN-PSO) method. The prediction was based on damage indices of 35 buildings around Malaysia. The BPNN-PSO demonstrated a better result of 89% accuracy compared to the traditional backpropagation neural network with only 84%. The capability of PSO supports fast convergence method has shown good effort to improve the processing time and accuracy of the results.

Estimating the rainfall threshold of a deep-seated landslide by integrating models for predicting the groundwater level and stability analysis of the slope

Rainfall thresholds of landslides often are determined by empirical meteorological thresholds, but the reliability of this approach sometimes is limited by the lack of information about the hydrological processes that trigger landslides. Groundwater plays a critical role in triggering deep-seated landslides. In this study, we propose a methodology to estimate the rainfall threshold for a deep-seated landslide based on an integrated model that combines a model for predicting the level of the groundwater with a finite-element, strength-reduction model. First, in order to obtain more accurate results when predicting the level of the groundwater, a method is proposed to estimate the groundwater level fluctuation caused by rainfall (GLFR). Then, two different machine learning methods, i.e., the genetic algorithm back-propagation neural network (GA-BPNN) method and the genetic algorithm support vector machine (GA-SVM) method, are proposed for predicting the GLFR of the Duxiantou landslide located in Zhejiang Province, China. The results of the predictions showed that the performance of the GA-SVM model was better than that of the GA-BPNN model. Then, a series of numerical simulations was conducted to investigate the factor of safety (Fs) of the slope at different groundwater levels. At last, the probabilities of the occurrences of Duxiantou landslides for different return periods of rainfall intensity were evaluated to determine the rainfall threshold.

Inspection routing problem for coal mine safety personnel in underground mines

Coal mine safety is the foundation of coal mine enterprise development, and inspections performed by safety personnel are an indispensable part of ensuring the production of coal mines. The complex network of tunnels in underground mines restricts the efficiency of safety inspection routes; therefore, planning these routes is critical. This study investigates a routing problem in which safety personnel start from multiple sites to inspect a set of sites. A mixed integer linear programming (MILP) model is formulated to optimize the inspection routes. Three methods are used to address the problem: a local branching (LB)-based solution method, a tabu search-based solution method and a back propagation (BP) neural network method. Numerical experiments are conducted to assess the efficiency of the methods on different scales.

Identification of SNM based on low-resolution gamma-ray characteristics and neural network

The risk of revealing sensitive information of nuclear weapon is an obstacle for comprehensively applying the identification technology in nuclear verification and nuclear security. In order to reduce the risk, low-resolution radiation spectra are suggested to be used in the activities of identifying special nuclear material (SNM) items’ types. In this article, we proposed an effective algorithm that extracts characteristic information from low-resolution gamma-ray spectra of SNMs and identifies the types of SNMs through backpropagation (BP) neural network and template matching method. We established the algorithm by numerical simulations, and then conducted series of experiments to verify and validate this algorithm. The identification results of applying this algorithm to real plutonium (PU) and high enriched uranium (HEU) pits showed that the proposed algorithm is an eligible option for both nuclear verification and nuclear security.

Expressway crash risk prediction using back propagation neural network: A brief investigation on safety resilience

This study presents the work in predicting crash risk on expressways with consideration of both the impact of safety critical events and traffic conditions. The traffic resilience theory is introduced to learn safety problems from the standpoint of 1) considering safety critical events, such as traffic violations, as the safety disturbances, and 2) considering safety resilience as the ability of the traffic, greatly associated with traffic conditions, to resist critical events turning into crashes. The concept of safety resilience was illustrated qualitatively through simulation experiments. Aimsun microsimulation software was used to simulate traffic conditions with safety critical events (vehicle violations, in this paper) involved based on the geometric design of the G15 Expressway in Shanghai. Based on data from the simulation experiment, a two-staged model was developed which classifies crash risk status into three types including no-risk, low-risk and high-risk status. Modeling approach that relies on the back propagation neural network method was applied. The performance of the model in prediction was validated through the Receiver Operating Characteristic (ROC) curve test. Results indicated that the model performed well in predicting crash risks in the simulated environment. After training the model, an extra simulation experiment involving six additional tests was conducted. Results show that the traffic resilience theory may work in explaining the relationship between traffic conditions, safety critical events and crash risk, which are the key elements in road safety field. The introduction of safety resilience may inspire further exploration on this topic in both research and practice. Meanwhile, the model can be used to predict and monitor risks on expressways in a potentially more precise way.

Forming Quality Control of an AA5182-O Aluminum Alloy Engine Hood Inner Panel

The poor formability of aluminum alloys at room temperature easily leads to quality defects such as ruptures, wrinkles and excessive springback after forming, which severely restricts their wide application in the automotive industry. In this article, a combination of springback compensation and process parameter optimization is proposed to improve the forming quality of aluminum alloy auto body panels. First, the springback compensation for the tooling setting was conducted using the global shape modeling (GSM) function in ThinkDesign to ensure the desired dimensional quality of the hood. Then, the process parameter optimization was conducted based on the combination of a back-propagation (BP) neural network and genetic algorithm (GA) method to improve formability. After springback compensation and process parameter optimization, the obtained product could satisfy the matching requirements well. A case study of an AA5182-O aluminum alloy engine hood inner panel is presented. The experiments demonstrate that the combination of the springback compensation and the optimization scheme based on the BP neural network and GA can effectively improve the product’s forming quality.

Feasibility analysis for acquiring visibility based on lidar signal using genetic algorithm-optimized back propagation algorithm**Project supported by the National Natural Science Foundation of China (Grant No. 41405014).

Visibility is an important atmospheric parameter that is gaining increasing global attention. This study introduces a back-propagation neural network method based on genetic algorithm optimization to obtain visibility directly using light detection and ranging (lidar) signals instead of acquiring extinction coefficient. We have validated the performance of the novel method by comparing it with the traditional inversion method, the back-propagation (BP) neural network method, and the Belfort, which is used as a standard value. The mean square error (MSE) and mean absolute percentage error (MAPE) values of the genetic algorithm-optimized back propagation (GABP) method are located in the range of 0.002 km2–0.005 km2 and 1%–3%, respectively. However, the MSE and MAPE values of the traditional inversion method and the BP method are significantly higher than those of the GABP method. Our results indicate that the proposed algorithm achieves better performance and can be used as a valuable new approach for visibility estimation.

Response prediction of multi-story building using backpropagation neural networks method

The active ground motion in Indonesia might cause a catastrophic collapse of the building which leads to casualties and property damages. Therefore, it is imperative to design the structural response of building against seismic hazard correctly. Seismic-resistant building design process requires structural analysis to be performed to obtain the necessary building responses. However, the structural analysis could be difficult and time-consuming. This study aims to predict the structural response includes displacement, velocity, and acceleration of multi-story building with the fixed floor plan using Backpropagation Neural Network (BPNN) method. By varying the building height, soil condition, and seismic location in 47 cities in Indonesia, 6345 datasets were obtained and fed into the BPNN model for the learning process. The trained BPNN is capable of predicting the displacement, velocity, and acceleration responses with up to 96% of the expected rate.

A Wireless Localization Algorithm With BPNN-MEA-QPSO for Connected Vehicle

As one of key technologies of connected vehicles (CVs) applications, wireless localization can provide accurate and reliable vehicle location for high occupancy tolling and safety critical vehicle applications, such as collision avoidance. Several artificial intelligence methods, such as back propagation neural network (BPNN) and particle swarm optimization (PSO) method, have been employed to optimize the pass-loss model and to improve the accuracy of wireless localization algorithm. However, in view of the stochasticity of initial weights and thresholds in BPNN, it is difficult to reach the global convergence. In this study, a novel double-layer architecture for wireless localization algorithm is proposed based on the optimization of initial weights and thresholds in BPNN and the refinement of search direction and step in PSO algorithm. Based on the architecture, the wireless localization algorithm integrating BPNN with mind evolutionary algorithm (MEA) and quantum-behaved PSO (QPSO) method is proposed and validated using the experiment data in field environment. The validation results show that the proposed localization algorithm has better localization accuracy by comparison with the other localization algorithms, which the average error of the proposed localization algorithm in field environment is about 19 meters. In addition, the localization accuracy shows an improving tendency with the increasing of the number of base stations connected to moving vehicle. The location error is less than 10 meters when the number of base stations connected to the moving vehicle is greater than 7. For the wireless localization in field environment, the accuracy is acceptable for CVs applications.

Stability and Complexity of a Novel Three‐Dimensional Environmental Quality Dynamic Evolution System

In this paper, a novel three‐dimensional environmental quality dynamic system is introduced. Bayesian estimation was used to calibrate environmental quality variables, and Genetic algorithm (GA) optimized Levenberg‐Marquardt Back Propagation (LM‐BP) neural network method was used to effectively identify the system parameters for calibration of various variables and official data. The studies found that the effect of increasing investment in environmental protection on energy intensity and environmental quality is not obvious, and it also aggravates the economic instability. Adjustment of peak parameters of pollution emissions can accelerate the evolution of energy intensity and environmental quality to a stable speed and eventually stabilize with a certain value. But if the peak value of pollution emissions reaches too early, it will pose a certain threat to the environment. Although the speed of ecological environment self‐repair is increased, it cannot effectively reduce energy intensity, improve environmental quality, and maintain economic growth; it can control the stability of the control system or effectively control pollution. Therefore, in order to improve the environmental quality, we need to take more measures in parallel, use more means and resources for environmental governance, and ultimately achieve “win‐win” between environmental quality and economy.

BP neural network prediction model for suicide attempt among Chinese rural residents.

This study aimed to establish and assess the Back Propagation Neural Network (BPNN) prediction model for suicide attempt, so as to improve the individual prediction accuracy. Data was collected from a wide range case-control suicide attempt survey. 659 serious suicide attempters (case group) were randomly recruited through the hospital emergency and patient registration system from 13 rural counties in China. Each case was matched the control by same community, gender, and similar age (± 2 ages). Face to face interviews were conducted for each subject with structured questionnaire. Logistic regression was applied to preliminarily screen the factors and BPNN was used to establish the prediction model of suicide attempt. Multivariate logistic regression indicated that family history of suicide (OR = 4.146), mental problem (OR = 3.876) Low education level, poor health, aspiration strain, hopelessness, impulsivity, depression are the risk predictors and social support, coping skills, healthy community are the protect predictors for suicide attempt. Repetitious data simulation process of BPNN indicated that three-layer BPNN with 9 hidden layer neurons is the optimal prediction model. The sensitivity (67.6%), specificity (93.9%), positive predictive value (86.0%), negative predictive value (84.1%), total coincidence rate (84.6%) manifested that it is excellent to distinguish suicide attempt case. The BPNN method is applicative, feasible, credible and good discriminative effect for suicide attempt. The BPNN established has significant clinical meaning for the clinical psychiatrist and lay theoretical foundation for artificial intelligence expert assisted diagnosis system for suicide attempt in the future.