Back-propagation Model Research Articles

Predicting the Mechanical Properties of Heat-Treated Woods Using Optimization-Algorithm-Based BPNN

This paper aims to enhance the accuracy of predicting the mechanical behavior of wood subjected to thermal modification using an improved dung beetle optimization (IDBO) model. The IDBO algorithm improves the original DBO algorithm via three main steps: (1) using piece-wise linear chaotic mapping (PWLCM) to generate the initial dung beetle species and increase its heterogeneity; (2) adopting an adaptive nonlinear decreasing producer ratio model to control the number of producers and boost the algorithm’s convergence rate; and (3) applying a dimensional learning-enhanced foraging (DLF) search strategy that optimizes the algorithm’s ability to explore and exploit the search space. The IDBO algorithm is evaluated on 14 benchmark functions and outperforms other algorithms. The IDBO algorithm is then applied to optimize a back-propagation (BP) neural network for predicting five mechanical property parameters of heat-treated larch-sawn timber. The results indicate that the IDBO-BP model significantly reduces the error compared with the BP, tent-sparrow search algorithm (TSSA)-BP, grey wolf optimizer (GWO)-BP, nonlinear adaptive grouping grey wolf optimizer (IGWO)-BP and DBO-BP models, demonstrating its superiority in predicting the physical characteristics of lumber after heat treatment.

Software Effort Estimation Based on Use Case Reuse (Back Propagation)

Abstract: Project managers often use effort estimating strategies to manage the human resources of current or upcoming software projects. Prior to project implementation, cost, time, andpersonnel estimation are basically necessary. For every project of software, getting accuracy in Effort Estimation has always been difficult. In this study, the estimation of software development effort was determined using a back propagation model. This model's goal is to investigate the capabilities and potential uses of Utilizing artificial neural networks (ANN) as a tool for forecasting the effort required for software development. In order to estimate the software work, we are attempting to implement a machine learning technique in this research. Out of all machine learning methods, we are applying an algorithm based on Artificial Neural Networks that is Back propagation. The Desharnais dataset, a well-known publicly available dataset for estimation of software effort, is used to test the approach. The performance and accuracy of the tested model have been evaluated using three metrics: MMRE, MRE, and Pred (0.25). In the sections below that follows, I explain the algorithm and its results

Estimation of Compressive Strengths of Pumice-and Diatomite-Containing Cement Mortars with Cascade Forward Back Propagation and Gaussian Function Models

Bu çalışmada, yapay sinir ağı (YSA) ve uyarlamalı ağ tabanlı bulanık çıkarım sistemi (ANFIS) ile pomza ve diyatomit içerikli çimento harçlarının basınç dayanımlarının tahmini yapılmıştır. YSA için kademeli ileri geri yayılım algoritması, ANFIS için ise Gauss üyelik fonksiyonu tercih edilmiştir. Modellerin oluşturulmasında toplam 7 tip çimento ile üretilen harçların 2., 7., 28. ve 90. hidratasyon günlerinde belirlenen basınç dayanım sonuçları kullanılmıştır. Modellerin eğitim ve test süreçlerinde; 5 giriş (hidratasyon günü, Portland çimento, pomza, diatomit, su) ve 1 çıkış (basınç dayanımı) parametresi kullanılmıştır. Deney sonuçlarıyla modelden elde edilen sonuçların karşılaştırılması R2, MAPE ve RMSE gibi istatistiksel yöntemlerle gerçekleştirilmiştir. Elde edilen veriler, YSA modeli ile elde edilen sonuçların tüm hidratasyon günleri için hemen hemen gerçek değerlere ulaşıldığını ve bu modelin başarılı bir tahmin modeli olduğunu göstermektedir.

Research on the Uncertainty of Landslide Susceptibility Prediction Using Various Data-Driven Models and Attribute Interval Division

Two significant uncertainties that are crucial for landslide susceptibility prediction modeling are attribute interval numbers (AIN) division of continuous landslide impact factors in frequency ratio analysis and various susceptibility prediction models. Five continuous landslide impact factor interval attribute classifications (4, 8, 12, 16, 20) and three data-driven models (deep belief networks (DBN), random forest (RF), and neural network (back propagation (BP)) were used for a total of fifteen different scenarios of landslide susceptibility prediction studies in order to investigate the effects of these two factors on modeling and perform a landslide susceptibility index uncertainty analysis (including precision evaluation and statistical law). The findings indicate that: (1) The results demonstrate that for the same model, as the interval attribute value rises from 4 to 8 and finally to 20, the forecast accuracy of landslide susceptibility initially increases gradually, then progressively grows until stable. (2) The DBN model, followed by the RF and BP models, provides the highest prediction accuracy for the same interval attribute value. (3) AIN = 20 and DBN models have the highest prediction accuracy under 15 combined conditions, while AIN = 4 and BP models have the lowest. The accuracy and efficiency of landslide susceptibility modeling are higher when the AIN = 8 and DBN models are combined. (4) The landslide susceptibility index uncertainty predicted by the deeper learning model and the bigger interval attribute value is comparatively low, which is more in line with the real landslide probability distribution features. The conditions that the environmental component attribute interval is divided into eight parts and DBN models are used allow for the efficient and accurate construction of the landslide susceptibility prediction model.

Characteristic mango price forecasting using combined deep-learning optimization model.

Accurate product price forecasting is helpful for scientific decision-making and precise industrial planning. As a characteristic fruit that drives regional development, mango price prediction is of great significance to several economies. However, owing to the strong volatility of mango prices, forecasting is vulnerable to uncertainties and is very challenging. In this study, a deep-learning combination forecasting model based on a back-propagation (BP) long short-term memory (LSTM) neural network is proposed. Using daily mango price data from a large fruit wholesale trading center in China from January 2nd, 2014, to April 18th, 2022, mango price changes are learned and predicted to support the fruit industry. The results show that the root mean-square error, mean absolute percentage error, and the R2 determination coefficient of the BP-LSTM combination model are 0.0175, 0.14%, and 0.9998, respectively. The prediction results of the combined model are better than those of the separate BP and LSTM models. Furthermore, it best fits the actual price profile and has better generalizability.

A Study of a Model for Predicting Pneumatic Subsoiling Resistance Based on Machine Learning Techniques

In order to explore the drag reduction mechanism of pneumatic subsoiling and study the influence of pneumatic subsoiling on the soil, this study used machine learning models to predict the working resistance of a pneumatic subsoiler and adopted random forest (RF), error back-propagation (BP), eXtreme gradient boosting (XGBoost) and support vector regression (SVR) to analyze and compare the predictions of these four models. Field experiments were carried out in two fields with different bulk densities and moisture content. The effects of these parameters on the resistance of pneumatic subsoiling were studied by changing the working air pressure, depth and forward speed. In the RF, SVR, XGBoost and BP models, five parameters (working air pressure, working depth, forward speed, bulk density and moisture content) were inputted as independent variables, and the operating resistance of pneumatic subsoiling was used as the predicted value. After training the four models, the results showed that the R2 value of the RF model was the highest and the error was the smallest, which made it better than the SVR, XGBoost and BP models. The values of MAPE, R2 and RMSE for the RF model’s test set were 0.01, 0.99, and 3.61 N, respectively, indicating that the RF model could predict the resistance value of subsoiling well. When the RF model was used to analyze the five input parameters, the experimental results showed that the contribution of working air pressure to reducing the resistance of subsoiling reached 29%, indicating that pneumatic subsoiling can reduce the resistance, drag and consumption.

Voids prediction beneath cement concrete slabs using a FEM-ANN method

ABSTRACT The voids beneath cement concrete slabs are a major invisible disease, resulting in a rapid decrease in service performance in the composite pavement. Accurate voids prediction is essential for the extensive application and long-term service of composite pavement. This research provides a FEM-ANN (Finite Element Modelling-Artificial Neural Network) method to predict the voids beneath concrete slabs. These ANN models include the original back propagation (BP), the particle swarm optimisation (PSO) BP model, the genetic algorithm (GA) BP model, and the whale optimisation algorithm (WOA) BP model. The voids FEM model is established and validated by the measured data in the field, and the relative error of measured and simulated results is within 4%. The cross-validation results show that the WOA-BP model has the best prediction performance, with the highest score of 8, which refers to the overall score of the mean value and variance of these evaluation indices. Therefore, this FEM-ANN framework is an efficient method for estimating the voids beneath concrete slabs. Furthermore, it is discovered that the base modulus with the highest contribution degree of 20.34% is the most dominant factor in predicting the voids output. HIGHLIGHTS A FEM-ANN method is utilised to predict the voids beneath concrete slabs The WOA-BP model exhibits the best comprehensive performance of the four ANN models. W d and pavement mechanical responses have a positive effect on A v opposite to K d and pavement structure. The base modulus is the primary factor in predicting the voids output.

Meta-learning biologically plausible plasticity rules with random feedback pathways

Backpropagation is widely used to train artificial neural networks, but its relationship to synaptic plasticity in the brain is unknown. Some biological models of backpropagation rely on feedback projections that are symmetric with feedforward connections, but experiments do not corroborate the existence of such symmetric backward connectivity. Random feedback alignment offers an alternative model in which errors are propagated backward through fixed, random backward connections. This approach successfully trains shallow models, but learns slowly and does not perform well with deeper models or online learning. In this study, we develop a meta-learning approach to discover interpretable, biologically plausible plasticity rules that improve online learning performance with fixed random feedback connections. The resulting plasticity rules show improved online training of deep models in the low data regime. Our results highlight the potential of meta-learning to discover effective, interpretable learning rules satisfying biological constraints.

Construction of an efficient evaluation model for athletic athletes' competitive ability based on deep neural network algorithm

This paper analyzes the data of this year's athletes' physical fitness test scores and manages the classification of different physical qualities of the farmer. In order to reduce the manual calculation and increase the prediction efficiency, as well as to unify the scoring criteria of previous years, this paper proposes a comprehensive performance prediction model based on deep neural network algorithm. First, principal component analysis is used to transform multiple attributes with strong correlation into independent attributes that are not related to each other, and to reduce the time and space for model training by eliminating redundancy. Second, a back propagation (BP) neural network algorithm is used to build a physical fitness test prediction model, and the model is applied to the test dataset for model performance evaluation. Finally, the physical fitness test model was applied to other years for comprehensive performance prediction, and the differences between the model prediction results and the actual teachers' manual calculation results were observed. The results showed very good prediction results for 2021, in which 92.95% of the data had an absolute value of error less than 2 and only 0.06% had an absolute value of error greater than 4, which indicated that the prediction performance of the model was extremely significant. At the same time, a new athletic athletic scoring standard was also developed based on the neural network BP model to provide a more scientific theoretical basis and guidance for the evaluation of athletic ability of athletes.

Physical Modeling and Intelligent Prediction for Instability of High Backfill Slope Moisturized under the Influence of Rainfall Disasters

The stability of high backfill slopes emerges in practice due to the expansion of transportation infrastructures. The seepage and infiltration of rainfall into the backfills brings challenges to engineers in predicting the stability of the slope, weakening the shear strength and modulus of the soil. This study carried out a series of model tests under a plane strain condition to investigate the stability of a high backfill slope moisturized by rainfalls, considering the influences of rainfall duration and intensity. The slope displacements were monitored by a laser displacement sensor and the moisture content in the backfill mass were obtained by a soil moisture sensor. The test results show that increasing the rainfall intensity and duration caused the slope near the surface to be saturated, resulting in significant influences on the lateral displacement of the slope and the reduction of stability as well as the sizes of the sliding mass. Based on the model tests, the numerical analysis was adopted to extend the analysis cases, and the backpropagation (BP) neural network model was further adopted to build a model for predicting the stability of a high backfill slope under rainfall. The trained BP model shows the average relative error of 1.02% and the goodness of fitness of 0.999, indicating a good prediction effect.

Nonstationary Time Series Prediction Based on Deep Echo State Network Tuned by Bayesian Optimization

The predictions from time series data can help us sense development trends and make scientific decisions in advance. The commonly used forecasting methods with backpropagation consume a lot of computational resources. The deep echo state network (DeepESN) is an advanced prediction method with a deep neural network structure and training algorithm without backpropagation. In this paper, a Bayesian optimization algorithm (BOA) is proposed to optimize DeepESN to address the problem of increasing parameter scale. Firstly, the DeepESN was studied and constructed as the basic prediction model for the time series data. Secondly, the BOA was reconstructed, based on the DeepESN, for optimal parameter searching. The algorithm is proposed within the framework of the DeepESN. Thirdly, an experiment was conducted to verify the DeepESN with a BOA within three datasets: simulation data generated from computer programs, a real humidity dataset collected from Beijing, and a power load dataset obtained from America. Compared with the models of BP (backpropagation), LSTM (long short-term memory), GRU (gated recurrent unit), and ESN (echo state network), DeepESN obtained optimal results, which were 0.0719, 18.6707, and 764.5281 using RMSE evaluation. While getting better accuracy, the BOA optimization time was only 323.4 s, 563.2 s, and 9854 s for the three datasets. It is more efficient than grid search and grey wolf optimizer.

Coordinated analysis and evaluation of water–energy–food coupling: A case study of the Yellow River basin in Shandong Province, China

Water resources, energy, and food are important basic resources for high-quality regional development. In the process of rapid development of regional economy, how to coordinate the development of basic resources has become one of the most serious challenges to the high-quality development of the Yellow River basin and even the whole Yellow River basin in Shandong province. Previous studies have produced few such analyses and evaluations for cities in the Yellow River basin, and there is a lack of research on multi-scale analysis, evaluation and prediction of urban coupling and coordinated development level. Hence, in this study, we propose an indicator system consisting of 24 indicators for the water resources, energy, and food system based on panel data for 9 cities in the Yellow River basin in Shandong Province for 2011–2020. Spatial autocorrelation analysis and temporal and spatial distribution analysis were used to determine the current state of water resources, energy, and food systems at multiple scales; combined weighting, comprehensive evaluation, and coupling coordination quantification were used to formulate the coupling coordination relationships of water resources–energy–food and identify the spatiotemporal formation and development of the water resources–energy–food system. A particle swarm optimization–back propagation model (PSO-BP) was used to predict the coupled and coordinated changes in the water resources–energy–food system for the next five years. The study results show that: (1) The composite index of the water resources–energy–food system showed an increasing trend, and the composite index of the food subsystem increased relatively quickly. (2) In 2020, the coupling coordination degree for eight of the nine cities in the Yellow River basin in Shandong Province was 0.70. Jinan, the exception, was at a critical stage of transitioning from low to moderate coupling, and the other eight cities were at the moderate stage of coupling coordination. (3) The coupling coordination of the water resources–energy–food system is increasing; it has transitioned from low to moderate and is about to transition to the stage of good coupling coordination. (4) The PSO-BP model predicted that the coupling coordination of water resources–energy–food systems in the Yellow River basin in Shandong Province will reach the stage of good coupling coordination in 2023. The research results presented in this paper can provide a theoretical basis for formulating policy recommendations in the Yellow River basin in Shandong Province.

Prediction of Microstructure and Mechanical Properties of Atmospheric Plasma-Sprayed 8YSZ Thermal Barrier Coatings Using Hybrid Machine Learning Approaches

The preparation of thermal barrier coatings (TBCs) is a complex process involving the integration of physics and chemistry, mainly involving the flight behavior and deposition behavior of molten particles. The service life and performance of the TBCs were determined by various factors, especially the preparation process parameters. In this work, to set up the quantitative characterization model between the preparation process parameters and the performance characteristic parameters, the ceramic powder particle size, spraying power and spraying distance were treated as the model input parameters, the characteristic parameters of microstructure properties represented by the porosity, circularity and Feret’s diameter and the mechanical property represented by the interfacial binding strength and macrohardness were treated as the model output. The typical back propagation (BP) model and extreme learning machine (ELM) model combined with flower pollination algorithm (FPA) optimization algorithm were employed for modeling analysis. To ensure the robustness of the obtained regression prediction model, the k-fold cross-validation method was employed to evaluate and analyze the regression prediction models. The results showed that the regression coefficient R value of the proposed FPA-ELM hybrid machine learning model was more than 0.94, the root-mean-square error (RMSE) was lower than 2 and showed better prediction accuracy and robustness. Finally, this work provided a novel method to optimize the TBCs preparation process, and was expected to improve the efficiency of TBCs preparation and characterization in the future.

Grade evaluation of black-odorous urban rivers in the Greater Bay Area of China using an improved back propagation (BP) neural network.

With the rapid development of urbanization, the urban water environment is receiving continuous attention. It is necessary to understand water quality in a timely manner and make a reasonable comprehensive evaluation. However, existing black-odorous water grade evaluation guidelines are not sufficient. Understanding the changing situation of black-odorous water in urban rivers is a growing concern, especially in real-world scenarios. In this study, a BP neural network combined with the fuzzy membership degree was applied to evaluate the black-odorous grade of urban rivers in Foshan City, which is within the Greater Bay Area of China. The optimal 4 × 11 × 1 topology structure of the BP model was constructed by taking the dissolved oxygen (DO), ammonia nitrogen (NH3-N), chemical oxygen demand (COD), and total phosphorus (TP) concentrations as input water quality indicators. There was almost no occurrence of black-odorous water in the two public rivers outside the region in 2021. Black-odorous water was most significant in 10 urban rivers, with grade IV and grade V occurring over 50% of the time in 2021. These rivers had three features, i.e., parallel with a public river, beheaded, and close proximity to Guangzhou City, the provincial capital of Guangdong. The results of the grade evaluation of the black-odorous water found basically matched those of the water quality assessment. The existence of some inconsistencies between the two systems justified the necessity to expand and extend the number of employed indicators and grades in the present guidelines. The results confirm the capability of the BP neural network combined with the fuzzy-based membership degree in the quantitative grade evaluation of black-odorous water in urban rivers. This study makes a step forward in understanding the grading of black-odorous urban rivers. The findings can provide a reference for local policy-makers regarding the priority of practical engineering projects in prevailing water environment treatment programs.

Optimization of BP Neural Network Model for Rockburst Prediction under Multiple Influence Factors

Rockbursts are serious threats to the safe production of mining, resulting in great casualties and property losses. The accurate prediction of rockburst is an important premise that influences the safety and health of miners. As a classical machine learning algorithm, the back propagation (BP) neural network has been widely used in rockburst prediction. However, there are few reports about the influence study of different training sample sizes, optimization algorithms and index dimensionless methods on the prediction accuracy of BP neural network models. Therefore, 100 groups of typical rockburst engineering samples were collected locally and abroad, and considering the relevance, scientificity and quantifiability of the prediction indexes, the ratio of the maximum tangential stress of surrounding rock to the rock uniaxial compressive strength (σθ/σc), the ratio of the rock uniaxial compressive strength to the rock uniaxial tensile strength (σc/σt) and the elastic energy index (Wet) were chosen as the prediction indexes. When the number of samples was 40, 70 and 100, sixty improved BP models were established based on the standard gradient descent algorithm and four optimization algorithms (momentum gradient descent algorithm, quasi-Newton algorithm, conjugate gradient algorithm, Levenberg–Marquardt algorithm) and four index dimensionless methods (unified extreme value processing method, differentiated extreme value processing method, data averaging processing method, normalized processing method). The prediction performances of each improved model were compared with those of standard BP models. The comparative study results indicate that the sample size, optimization algorithm and dimensionless method have different effects on the prediction accuracy of BP models, which are described as follows: (1) The prediction accuracy value A of the BP model increases with the addition of sample size. The average value Aave of twenty improved models under three kinds of sample sizes increases from Aave (40) = 69.7% to Aave (100) = 75.3%, with a maximal value Amax from Amax (40) = 85.0% to Amax (100) = 97.0%. (2) The value A and comprehensive accuracy value C of the BP model based on four optimization algorithms are generally higher than those of the standard BP model. (3) The improved BP model based on the unified extreme value processing method combined with the Levenberg–Marquardt algorithm has the highest value Amax (100) = 97.0% and value C = 194, and the prediction results of five engineering cases are completely consistent with the actual situation at the site, so this is the best BP neural network model selected in this paper.

Evaluation of thermal comfort in air-conditioned rooms based on structure/control-related parameters and data-mining method

Evaluating the thermal environment and thermal comfort in an air-conditioned room is an essential for estimating the performance of air-conditioning systems. However, multiple component structures and control-related parameters often lead to a long test cycle and large number of tests, significantly affecting the testing efficiency and speed. To address these problems, in this study, a data-mining method was proposed to predict and evaluate the thermal environment of an air-conditioned room. Owing to the limited amount of experimental data, the sample data were expanded by the simulation data of a collaborative platform between the air-conditioning system and air-conditioned room. Data-mining models, including the support vector regression (SVR), backpropagation (BP), and multiple linear regression (MLR) models, were developed and achieved good accuracy in evaluating the thermal environment by considering air-conditioning systems with various structures and control parameters. In the multiple-input single-output evaluation method, the prediction accuracy of the SVR model was higher than those of the BP and MLR models with respect to the vertical air temperature difference, temperature uniformity, temperature drop rate, and draft rate, while the result was the opposite in terms of the predicted mean vote indices. In the multiple-input multiple-output evaluation method, there was a decline in prediction accuracy and an increase in efficiency prediction compared with multiple-input single-output evaluation.

Mind evolutionary algorithm optimization in the prediction of satellite clock bias using the back propagation neural network

Satellite clock bias is the key factor affecting the accuracy of the single point positioning of a global navigation satellite system. The traditional model back propagation (BP) neural network is prone to local optimum problems. This paper presents a prediction model and algorithm for the clock bias of the BP neural network based on the optimization of the mind evolutionary algorithm (MEA), which is used to optimize the initial weights and thresholds of the BP neural network. The accuracy of the comparison between clock bias data is verified with and without one-time difference processing. Compared with grey model (GM (1,1)) and BP neural network, this paper discusses the advantages and general applicability of this method from different constellation satellites, different atomic clock type satellites, and the amount of modeling data. The accuracy of the grey model (GM(1,1)), BP, and MEA-BP models for satellite clock bias prediction is analyzed and the root mean square error, range difference error, and the mean of the clock bias data compared. The results demonstrate that the prediction accuracy of the three satellites significantly increased after one-time difference processing and that they have good stability. The prediction accuracy of four sessions of 2 h, 3 h, 6 h, and 12 h obtained using the MEA-BP model was better than 0.74, 0.80, 1.12, and 0.87 ns, respectively. The MEA-BP model has a specific degree of improvement in the prediction accuracy of the different sessions. Additionally, the prediction accuracy of different models has a specific relationship with the length of the original modeling sequence, of which BP model is the most affected, and MEABP is relatively less affected by the length of the modeling sequence, indicating that the MEA-BP model has strong anti-interference ability.

Landslide Displacement Prediction of Shuping Landslide Combining PSO and LSSVM Model

Predicting the deformation of landslides is significant for landslide early warning. Taking the Shuping landslide in the Three Gorges Reservoir area (TGRA) as a case, the displacement is decomposed into two components by a time series model (TSM). The least squares support vector machine (LSSVM) model optimized by particle swarm optimization (PSO) is selected to predict the landslide displacement prediction based on rainfall and reservoir water level (RWL). Five parameters, including rainfall over the previous month, rainfall over the previous two months, RWL, change in RWL over the previous month and period displacement over the previous half year, are selected as the input variables. The relationships between the five parameters and the landslide displacement are revealed by grey correlation analysis. The PSO-LSSVM model is used to predict the periodic term displacement (PTD), and the least squares method is applied to predict the trend term displacement (TTD). With the same input variables, the back propagation (BP) model and the PSO-SVM model are also developed for comparative analysis. In the PSO-LSSVM model, the R2 of three monitoring stations is larger than 0.98, and the MAE values and the RMSE values are the smallest among the three models. The outcomes demonstrate that the PSO-LSSVM model has a high accuracy in predicting landslide displacement.

Regional/Single Station Zenith Tropospheric Delay Combination Prediction Model Based on Radial Basis Function Neural Network and Improved Long Short-Term Memory

Atmospheric water vapor is an essential source of information that predicts global climate change, rainfall, and disaster-natured weather. It is also a vital source of error for Earth observation systems, such as the global navigation satellite system (GNSS). The Zenith Tropospheric Delay (ZTD) plays a crucial role in applications, such as atmospheric water vapor inversion and GNSS precision positioning. ZTD has specific temporal and spatial variation characteristics. Real-time ZTD modeling is widely used in modern society. The conventional back propagation (BP) neural network model has issues, such as local, optimal, and long short-term memory (LSTM) model needs, which help by relying on long historical data. A regional/single station ZTD combination prediction model with high precision, efficiency, and suitability for online modeling was proposed. The model, called K-RBF, is based on the machine learning algorithms of radial basis function (RBF) neural network, assisted by the K-means cluster algorithm (K-RBF) and LSTM of real-time parameter updating (R-LSTM). An online updating mechanism is adopted to improve the modeling efficiency of the traditional LSTM. Taking the ZTD data (5 min sampling interval) of 13 international GNSS service stations in southern California in the United States for 90 consecutive days, K-RBF, R-LSTM, and K-RBF were used for regions, single stations, and a combination of ZTD prediction models regarding research, respectively. Real-time/near real-time prediction results show that the root-mean-square error (RMSE), mean absolute error (MAE), coefficient of determination (R2), and training time consumption (TTC) of the K-RBF model with 13 station data are 8.35 mm, 6.89 mm, 0.61, and 4.78 s, respectively. The accuracy and efficiency of the K-RBF model are improved compared with those of the conventional BP model. The RMSE, MAE, R2, and TTC of the R-LSTM model with WHC1 station data are 6.74 mm, 5.92 mm, 0.98, and 0.18 s, which improved by 67.43%, 66.42%, 63.33%, and 97.70% compared with those of the LSTM model. The comparison experiments of different historical observation data in 24 groups show that the real-time update model has strong applicability and accuracy for the time prediction of small sample data. The RMSE and MAE of K-RBF with 13 station data are 4.37 mm and 3.64 mm, which improved by 47.70% and 47.20% compared to K-RBF and by 28.48% and 31.29% compared to R-LSTM, respectively. The changes in the temporospatial features of ZTD are considered, as well, in the combination model.

Kajian Model Backpropagation dan Hybrid ANFIS Dalam Memprediksi Pertumbuhan Penduduk di Kabupaten Karawang

Population growth rate prediction is a process of estimating the population in the future. Predictions are made so that the government can prepare strategic steps in anticipating the negative impact of an uncontrolled population increase. The research data is the population of Karawang Regency from 2011 to 2020. Backpropagation and Hybrid ANFIS are the models used in this study. The purpose of this study was to determine the RMSE value and scatter data formed from the results of the ANFIS Backpropagation and Hybrid training models in predicting population growth rates in Karawang Regency. In addition, this study is intended to determine the level of accuracy of the two models. The research step begins with research data validation, preprocessing, training and testing, as well as accuracy testing. Accuracy testing uses the Mean Absolute Percentage Error (MAPE) method. Backpropagation and Hybrid models in predicting the rate of population growth have worked well. This can be seen from the training results of the two models. Backpropagation model has the best RMSE of 0.0328 and Hybrid has the best RMSE of 0.021884. The results of the analysis of the accuracy of predicting population growth rates for 2019 and 2020 that have been carried out, both models have a good level of accuracy. Backpropagation has an average accuracy rate of 84.76%, while the Hybrid model has an average accuracy rate of 93.71%. Based on the results of accuracy testing, the Hybrid model has a better level of accuracy than the Backpropagation model.