LSSVM Models Research Articles

Estimation Model for Maize Multi-Components Based on Hyperspectral Data.

Assessing the quality of corn seeds necessitates evaluating their water, fat, protein, and starch content. This study integrates hyperspectral imaging technology with chemometric analysis techniques to achieve non-invasive and rapid detection of multiple key components in corn seeds. Hyperspectral images of the embryo surface of maize seeds were collected within the wavelength range of 1100~2498 nm. Subsequently, image segmentation techniques were applied to extract the germ structure of the corn seeds as the region of interest. Seven spectral data preprocessing algorithms were employed, and the Detrending Transformation (DT) algorithm was identified as the optimal preprocessing method through comparative analysis using the Partial Least Squares Regression (PLSR) model. To reduce spectral redundancy and streamline the prediction model, three algorithms were employed for characteristic wavelength extraction: Successive Projections Algorithm (SPA), Competitive Adaptive Reweighted Sampling (CARS), and Uninformative Variable Elimination (UVE). Using the original spectra and extracted characteristic wavelengths, PLSR, BP, RBF, and LSSVM models were constructed to detect the content of four components. The analysis indicated that the CARS-LSSVM algorithm had the best prediction performance. The PSO algorithm was employed to further optimize the parameters of the LSSVM model, thereby improving the model's prediction performance. The R values for the four components in the test set were 0.9884, 0.9490, 0.9864, and 0.9687, respectively. This indicates that hyperspectral technology combined with the DT-CARS-PSO-LSSVM algorithm can effectively detect the main component content of corn seeds. This study not only provides a scientific basis for the evaluation of corn seed quality but also opens up new avenues for the development of non-destructive testing technology in related fields.

A novel grey seasonal model with time power for energy prediction

The fluctuation of seasonal time series has attracted considerable attention. However, the complexity of socio-economic development and the variability of influencing factors make accurate predictions of seasonal time series more difficult. A new discrete grey seasonal model, namely DGSTPM(1,1), is established, which employs the time power term based on the cultural algorithm to effectively deal with the time sequences characterized by nonlinear tendency and periodic fluctuations. Firstly, the operating mechanism of the DGSTPM(1,1) model is expounded by combining the time power item with the traditional grey seasonal model to enhance the forecasting capability. Then, the optimal time power index of the model can be determined by incorporating the cultural algorithm. Subsequently, several properties are thoroughly discussed to illustrate the superiority and uniformity of the new model. Finally, two energy-related cases with different data features and sample sizes are applied to identify the efficacy and robustness of the DGSTPM(1,1) model, referring to monthly crude oil production in the United States and quarterly coke production in China. Based on the various error criteria, SPA and DM tests, the results indicate that the DGSTPM(1,1) outperforms its competitors both in simulation and prediction phases, including DGSTM(1,1), SGM(1,1), SDTGM, FTGM, SARIMA, LSSVM, BPNN, and LSTM models. Meanwhile, the new model can effectively address seasonal time series with nonlinear characteristics owing to its dynamic time power item with the support of the cultural algorithm.

Study on wind power prediction based on improved double wavelet transform and quantile regression forest

Abstract The random fluctuation of a wind power system will bring significant challenges to its stable operation when connected to the grid[1]. To reduce the impact of microgrid wind power integration on electric power systems and in response to the certain prediction inaccuracy of BP, LSSVM, and ARIMA-based prediction models in case of rapid fluctuation of actual power, a method of decomposing the original wind power sequence into several characteristic subcomponents through the improved double wavelet transform algorithm was proposed in this study to weaken the fluctuation of the wind power sequence. Specifically, a short-term wind power probability density prediction model based on the improved double wavelet algorithm and quantile regression method was established. Then, the original wind power sequence was decomposed into a series of components with different frequencies using the double wavelet transform method. Suitable components were chosen to construct a QRF prediction model to acquire the prediction results.

A Single-Buyer Model of Imbalance Cost Pass-Through Pricing Forecasting in the Malaysian Electricity Supply Industry

The imbalance cost pass-through (ICPT) is a flexible component of the incentive-based regulation (IBR) that empowers power producers to adjust tariffs in response to variable fuel prices, thereby enhancing the economic resilience of electricity generation. In Malaysia, the Energy Commission has conducted biannual reviews of fuel and other generation costs. Any cost savings or increases identified during these reviews will be passed on to customers in the form of rebates or surcharges. Meanwhile, if an increment in the ICPT price signal can be provided to electricity providers and consumers, early preparation for operation budgeting can be realised, and energy management program development can be properly prepared. Due to this reason, this study proposes ICPT price forecasting for the electricity market in Peninsular Malaysia that will benefit the stakeholders. The study aims to construct an ICPT-related baseline model for the peninsular generation data by employing three forecasting methods. The forecasting performance is analysed using the mean absolute percentage error (MAPE). In light of our findings, the ARIMA method is one of the most accurate forecasting methods for fuel prices compared to the moving average (MA) and LSSVM methods. The observed price differences between the ARIMA and LSSVM models for ICPT are minimal. The ICPT price for July–December 2022 and January–June 2023 is MYR 0.21/kWh for the ARIMA and MYR 0.18/kWh for LSSVM, which are close to the actual TNB’s ICPT tariff. As for forecasting, the ICPT price is expected to drop in the next announcement. The findings of this study may have a positive impact on the sustainability of the energy sector in Malaysia.

Experimental analysis and model prediction of elbow pipe's erosion in water-cooled radiator

The radiator with heat transfer capability is able to guarantee the stable operation of hydro generator set, while the long-term and continuous scouring on radiator pipes by cooling medium will lead to thinning or even perforation of pipe wall, which triggers wall failure. This paper analyzes and predicts the failure mechanism of radiator’s pipe wall, and investigates the effects of water flow velocity, sand content and sand particle size on erosion damage of radiator pipe by establishing a test bench for pipe erosion. The results show that the increase of above parameters will lead to the increasing erosion rate, especially when the sand content is 1%, the velocity is 8 m/s and the sand particle size is 0.85 mm, the erosion damage will be particularly serious. Based on experimental data, BP and LSSVM models are employed to predict the pipe wall failure, and PSO algorithm is used to optimize the two models. The optimized PSO-BP has the highest accuracy with the mean absolute error (MAE) of 0.2070 and the mean absolute percentage error (MAPE) of 4.702%. The findings provide a reference for wall failure analysis of radiator, which is of great significance for unit's safe operation.

Identification of Risk Factors Associated with Tuberculosis in Southwest Iran: A Machine Learning Method.

Tuberculosis is a principal public health issue. Reducing and controlling tuberculosis did not result in the expected success despite implementing effective preventive and therapeutic programs, one of the reasons for which is the delay in definitive diagnosis. Therefore, creating a diagnostic aid system for tuberculosis screening can help in the early diagnosis of this disease. This research aims to use machine learning techniques to identify economic, social, and environmental factors affecting tuberculosis. This case-control study included 80 individuals with TB and 172 participants as controls. During January-October 2021, information was collected from thirty-six health centers in Ahvaz, southwest Iran. Five different machine learning approaches were used to identify factors associated with TB, including BMI, sex, age , marital status, education, employment status, size of the family, monthly income, cigarette smoking, hookah smoking, history of chronic illness, history of imprisonment, history of hospital admission, first-class family, second-class family, third-class family, friend, co-worker, neighbor, market, store, hospital, health center, workplace, restaurant, park, mosque, Basij base, Hairdressers and school. The data was analyzed using the statistical programming R software version 4.1.1. According to the calculated evaluation criteria, the accuracy level of 5 SVM, RF, LSSVM, KNN, and NB models is 0.99, 0.72, 0.97,0.99, and 0.95, respectively, and except for RF, the other models had the highest accuracy. Among the 39 investigated variables, 16 factors including First-class family (20.83%), friend (17.01%), health center (41.67%), hospital (24.74%), store (18.49%), market (14.32%), workplace (9.46%), history of hospital admission (51.82%), BMI (43.75%), sex (40.36%), age (22.83%), educational status (60.59%), employment status (43.58%), monthly income (63.80%), addiction (44.10%), history of imprisonment (38.19%) were of the highest importance on tuberculosis. The obtained results demonstrated that machine-learning techniques are effective in identifying economic, social, and environmental factors associated with tuberculosis. Identifying these different factors plays a significant role in preventing and performing appropriate and timely interventions to control this disease.

A novel machine learning-based imputation strategy for missing data in step-stress accelerated degradation test

The presence of missing data is a significant data quality issue that negatively impacts the accuracy and reliability of data analysis. This issue is especially relevant in the context of accelerated tests, particularly for step-stress accelerated degradation tests. While missing data can occur due to objective factors or human error, high missing rate is an inevitable pattern of missing data that will occur during the conversion process of accelerated test data. This type of missing data manifests as a degradation dataset with unequal measuring intervals. Therefore, developing a more appropriate imputation method for accelerated test data is essential. In this study, we propose a novel hybrid imputation method that combines the LSSVM and RBF models to address missing data problems. A comparison is conducted between the proposed model and various traditional and machine learning imputation methods using simulation data, to justify the advantages of the proposed model over the existing methods. Finally, the proposed model is implemented on real degradation datasets of the super-luminescent diode (SLD) to validate its performance and effectiveness in dealing with missing data in step-stress accelerated degradation test. Additionally, due to the generalizability of the proposed method, it is expected to be applicable in other scenarios with high missing data rates.

Determination of N-Paraffins Content in Crude Oil via Near-Infrared Spectroscopy Associated with Chemometric Approaches

This study explores the potential application of NIR spectroscopy coupled with different linear and nonlinear models for rapid evaluation of n-alkanes in crude oil. Samples for calibration were 30 model mixtures of n-eicosane in crude oil samples with a concentration of 1–15%. The prediction models were established based on 21 methods: linear regression, regression trees, support vector machines, Gaussian process regression, ensembles of trees, and neural networks. The spectral range 4500–9000 cm−1 was determined to be the most informative for prediction. The prediction capability of lineal regression methods turned out to be unsatisfactory. Nonlinear models were preferred over linear models; better results were obtained using the regression trees method, including «fine tree» (RMSE = 2.8635) and neural networks (RMSE = 2.0157). The LS-SVM model exhibited satisfactory prediction performance (R2 = 0.96, RMSE = 0.91), as did the Gaussian Process Regression Matern 5.2 GPR (R2 = 0.96, RMSE = 1.03) and Gaussian Process Regression (Rational Quadratic) (R2 = 0.95, RMSE = 1.04). Among the 21 chemometric algorithms, the best and weakest models were the LS-SVM and PLSR models, respectively. The LS-SVM model was the optimal model for the prediction of n-alkanes content in crude oil.

Early bruising detection of 'Korla' pears by low-cost visible-LED structured-illumination reflectance imaging and feature-based classification models.

Nondestructive detection of thin-skinned fruit bruising is one of the main challenges in the automated grading of post-harvest fruit. The structured-illumination reflectance imaging (SIRI) is an emerging optical technique with the potential for detection of bruises. This study presented the pioneering application of low-cost visible-LED SIRI for detecting early subcutaneous bruises in 'Korla' pears. Three types of bruising degrees (mild, moderate and severe) and ten sets of spatial frequencies (50, 100, 150, 200, 250, 300, 350, 400, 450 and 500 cycles m-1) were analyzed. By evaluation of contrast index (CI) values, 150 cycles m-1 was determined as the optimal spatial frequency. The sinusoidal pattern images were demodulated to get the DC, AC, and RT images without any stripe information. Based on AC and RT images, texture features were extracted and the LS-SVM, PLS-DA and KNN classification models combined the optimized features were developed for the detection of 'Korla' pears with varying degrees of bruising. It was found that RT images consistently outperformed AC images regardless of type of model, and LS-SVM model exhibited the highest detection accuracy and stability. Across mild, moderate, severe and mixed bruises, the LS-SVM model with RT images achieved classification accuracies of 98.6%, 98.9%, 98.5%, and 98.8%, respectively. This study showed that visible-LED SIRI technique could effectively detect early bruising of 'Korla' pears, providing a valuable reference for using low-cost visible LED SIRI to detect fruit damage.

Electrochemical degradation of ciprofloxacin from water: Modeling and prediction using ANN and LSSVM

Ciprofloxacin is a widely used antibiotic that is also a persistent contamination that causes major health and environmental risks. This study investigated the electrochemical removal of ciprofloxacin using a boron-doped diamond anode and a non-supportive electrode to get an accurate understanding of the anode's performance. The effects of current density, initial ciprofloxacin concentration, time, pH, and electrolyte concentration on the removal efficiency of ciprofloxacin were investigated. Additionally, ANN and LSSVM models were used to predict ciprofloxacin removal. The results showed that the BDD anode was effective in removing ciprofloxacin from water, with complete removal of 10 mg L−1 CIP in less than 45 min with electrolyte concentration of 25 g L−1, a current density of 20 mA cm−2, and a pH value of 3. The removal efficiency was found to increase with increasing current density and time because it increases the concentration of active species. The removal efficiency was also found to be higher at acidic pH values and higher electrolyte concentrations. The ANN model was found to be more accurate in predicting ciprofloxacin removal than the LSSVM model, with an R2 of 0.9982 and an AARE% of 2.68%, compared to R2 of 0.9978 and an AARE% of 3.85%. The models were validated using a new set of data, and the results showed that the ANN model was able to accurately predict ciprofloxacin removal under a variety of conditions.

Detection of moisture content and size of pumpkin seeds based on hyperspectral reflection and transmission imaging techniques

Moisture content and size can affect the vigor and weight of pumpkin seeds, where vigor is the key to evaluating seed quality, and weight is an important characteristic of the seeds. Therefore, detecting the moisture content and size of pumpkin seeds contributes to improving seed quality. In this study, hyperspectral reflection and transmission imaging techniques were used to detect the moisture and size of single pumpkin seeds. Linear PLSR and nonlinear LSSVM models were established to predict the moisture content of single pumpkin seeds using reflection and transmission spectral data. Five variable selection and data fusion methods, namely randomization test, variable selection based on C-value, uninformative variable elimination, Monte Carlo-uninformative variable elimination, and competitive adaptive reweighted sampling (CARS), were adopted to optimize the models. Compared with the models based on reflection and transmission spectra, low-level data fusion improved the model performance. The mid-level data fusion with models based on the CARS algorithm achieved the optimal performance, corresponding to R2P and RMSEP of 0.9219, 0.0279%, and 0.9231, 0.0278%, for the PLSR and LSSVM models, respectively. Finally, it was demonstrated that using hyperspectral reflection images to determine the size of pumpkin seeds is feasible, and the length and width of all samples were determined. The results show that hyperspectral imaging techniques and data fusion can effectively detect the moisture content of single pumpkin seeds, and measuring the size of pumpkin seeds is feasible using image processing algorithms.

Global-local least-squares support vector machine (GLocal-LS-SVM).

This study introduces the global-local least-squares support vector machine (GLocal-LS-SVM), a novel machine learning algorithm that combines the strengths of localised and global learning. GLocal-LS-SVM addresses the challenges associated with decentralised data sources, large datasets, and input-space-related issues. The algorithm is a double-layer learning approach that employs multiple local LS-SVM models in the first layer and one global LS-SVM model in the second layer. The key idea behind GLocal-LS-SVM is to extract the most informative data points, known as support vectors, from each local region in the input space. Local LS-SVM models are developed for each region to identify the most contributing data points with the highest support values. The local support vectors are then merged at the final layer to form a reduced training set used to train the global model. We evaluated the performance of GLocal-LS-SVM using both synthetic and real-world datasets. Our results demonstrate that GLocal-LS-SVM achieves comparable or superior classification performance compared to standard LS-SVM and state-of-the-art models. In addition, our experiments show that GLocal-LS-SVM outperforms standard LS-SVM in terms of computational efficiency. For instance, on a training dataset of 9, 000 instances, the average training time for GLocal-LS-SVM was only 2% of the time required to train the LS-SVM model while maintaining classification performance. In summary, the GLocal-LS-SVM algorithm offers a promising solution to address the challenges associated with decentralised data sources and large datasets while maintaining high classification performance. Furthermore, its computational efficiency makes it a valuable tool for practical applications in various domains.

Establishment of a multi-position general model for evaluation of watercore and soluble solid content in ‘Fuji’ apples using on-line full-transmittance visible and near infrared spectroscopy

Watercore and soluble solid content (SSC) are important indicators for assessment of the internal quality of apples. Visible and near infrared spectroscopy (Vis-NIRS) has been successfully applied for on-line analysis of internal quality of apples. In practical application, the position of the detected apples is usually uncertain, which limits the effective application of traditional model developed based on the fixed position or the optimal position. This study established a multi-position general model for evaluation of watercore and SSC in ‘Fuji’ apples using on-line Vis-NIRS. First, the full-transmittance spectra of three positions (P1–P3) of each apple were acquired, and subsequently, LS-SVM and PLS-DA models were constructed based on full wavelength data of single and multiple positions for watercore classification, and LS-SVM and PLS models were also constructed based on the same data for the quantitative prediction of SSC. Then, the detection ability of the constructed single position and the multi-position general models for three independent position samples was compared to determine the optimal full wavelength models; Finally, different wavelength selection algorithms including MC-UVE, BOSS, SPA and their combinations were employed to optimize the full wavelength models. Results showed that the multi-position general MC-UVE-SPA-PLS-DA model was the optimal for classification of watercore apples with average accuracy of 95.96 %, and the multi-position general MC-UV-BOSS-LS-SVM model was the optimal for quantitative detection of SSC with RP of 0.856 and RMSEP of 0.723 %. The overall study revealed that the multi-position general model, combined with full-transmittance Vis-NIRS, was a better choice for on-line detection of watercore and SSC in ‘Fuji’ apples.

Short-term wind power forecasting and uncertainty analysis based on FCM–WOA–ELM–GMM

With large-scale wind power connected to the power grid, accurate short-term wind power forecasting has become a key technology for safe, economic power grid operation. Therefore, a short-term wind power forecasting and uncertainty analysis method based on the FCM-WOA-ELM-GMM was proposed. Fuzzy C-means (FCM) was used to cluster the numerical weather prediction (NWP) and wind farm power data, and the data points with similar meteorological information are classified into one class. Using the rapid convergence and high convergence accuracy of the whale optimization algorithm (WOA), the input weight and hidden layer threshold of the extreme learning machine (ELM) model were optimized to improve the ELM calculation speed and forecasting accuracy. The WOA-ELM model was trained using the clustered NWP and wind farm power data and the short-term wind power was projected using the trained forecasting model. To accurately calculate the forecasting error probability density distribution, the Gaussian mixture model (GMM) was applied and the wind power forecast confidence intervals under different climatic conditions and time scales were calculated. The forecasting accuracies of the WOA-ELM, ELM, PSO-LSSVM, LSSVM, LSTM, PSO-BP, and WNN models were compared and analyzed, and the RMSE values of the 4-h forecasting results in April were as follows: WOA-ELM, 5.95%; ELM, 26.73%; PSO-LSSVM, 3.78%; LSSVM, 5.19%; LSTM, 23.71%; PSO-BP, 15.63%; and WNN, 23.23%. The RMSE values of 24-h forecasting results in April were: WOA-ELM, 6.62%; ELM, 19.86%; PSO-LSSVM, 9.91%; LSSVM, 13.73%; LSTM, 23.69%; PSO-BP, 14.08%; and WNN, 20.11%. The RMSE values of 72-h forecasting results in April were as follows: WOA-ELM, 5.24%; ELM, 13.64%; PSO-LSSVM, 12.03%; LSSVM, 13.67%; LSTM, 16.61%; PSO-BP, 15.46%; and WNN, 20.22%. According to the calculation results, under different climatic conditions and forecasting time scales, the forecasting accuracy of the FCM-WOA-ELM-GMM model is higher than those of the other models.

Designing a Multi-Stage Expert System for daily ocean wave energy forecasting: A multivariate data decomposition-based approach

Accurate forecasting of the wave energy is crucial and has significant potential because every wave meter possesses an energy amount ranging from 30 to 40 kW along the shore. By harnessing, it does not produce toxic gases, which is a better alternative to the energies that use fossil fuels. In this research, a multi-stage Multivariate Variational Mode Decomposition (MVMD) integrated with Boruta-Extreme Gradient Boosting (BXGB) feature selection and Cascaded Forward Neural Network (CFNN) (i.e., MVMD-BXGB-CFNN) is proposed to forecast daily ocean wave energy in the regions of Queensland State, Australia. The modelling outcomes were benchmarked via three other robust intelligence-based alternatives comprised of Multigene Genetic Programming (MGGP), Least Square Support Machine (LSSVM), and Gradient Boosted Decision Tree (GBDT) models hybridized with MVMD and BXGB (i.e., MVMD-BXGB-MGGP, MVMD-BXGB-LSSVM, and MVMD-BXGB-GBDT), and their counterpart standalone CFNN, GBDT, LSSVM, and MGGP models. To develop the multi-step hybrid intelligent systems, first, the primary input signals were simultaneously decomposed into intrinsic mode functions (IMFs) and residual components using the MVMD pre-processing technique. Next, the significant lags at the t-1 and t-2 timescales computed using the cross-correlation function were imposed on the decomposed components and further filtered by the BXGB feature selection to identify the best IMFs and reduce the computational cost and enhance the accuracy. Finally, the filtered IMFs were incorporated into the machine learning (ML) models to forecast the wave energy. Forecasting performance of all the provided models (hybrid and counterpart standalone ones) was evaluated during the testing phase by several well-known metrics, infographic tools, and diagnostic analysis. The results showed that the MVMD-BXGB-CFNN technique, as a capable expert system, outperformed the other hybrid and counterpart standalone methods and has an adequate degree of reliability to forecast the daily wave energy in coastal regions.

Fusion of spectra and texture data of hyperspectral imaging for prediction of myoglobin content in nitrite-cured mutton

The amount and chemical state of myoglobin are often relevant to meat color, which is one of the important characteristics of meat freshness and wholesomeness. To precisely determine OxyMb and MetMb contents in mutton, the feasibility of optimization detection technology of spectral image fusion was explored. Three methods were used to select characteristic wavelengths while textural features of muscle tissue were derived. Compared among all models, LSSVM models coupled with spectra and textural parameters all performed better than these PLSR models. The iVISSA-COR-LSSVM model for OxyMb prediction achieved a good performance of R2C = 0.9525, RMSEC = 2.2955, R2P = 0.8956, and RMSEP = 3.0036. The more effective prediction performances emerged from VCPA-IRIV-COR-LSSVM model with R2C = 0.8809, RMSEC = 2.2009, R2P = 0.8353, and RMSEP = 3.2096 for MetMb. In final, the pseudo color maps of OxyMb and MetMb distribution were generated. This study demonstrated that the data fusion method furnishes a new approach to rapidly evaluate myoglobin content.

Intelligent Modeling of the Incineration Process in Waste Incineration Power Plant Based on Deep Learning

The incineration process in waste-to-energy plants is characterized by high levels of inertia, large delays, strong coupling, and nonlinearity, which makes accurate modeling difficult. Therefore, an intelligent modeling method for the incineration process in waste-to-energy plants based on deep learning is proposed. First, the output variables were selected from the three aspects of safety, stability and economy. The initial variables related to the output variables were determined by mechanism analysis and the input variables were finally determined by removing invalid and redundant variables through the Lasso algorithm. Secondly, each delay time was calculated, and a multi-input and multi-output model was established on the basis of deep learning. Finally, the deep learning model was compared and verified with traditional models, including LSSVM, CNN, and LSTM. The simulation results show that the intelligent model of the incineration process in the waste-to-energy plant based on deep learning is more accurate and effective than the traditional LSSVM, CNN and LSTM models.

Estimating mutual solubility of a CO2 in NaCl aqueous solution system using connectionist approaches

Expert models of GP, MLP-ANNs and LSSVM were used to evaluate CO2 solubility in NaCl aqueous solution. Temperature, pressure, and molarity of NaCl aqueous solution were considered as the inputs of the model and CO2 (mole/kg) solubility in NaCl solutions was the models output. The proposed GP, LSSVM, and ANNs models are verified by comparing their results with extensive experimental data for CO2–NaCl solutions solubility from open literature. LSSVM modeling approach can be applied to model the system with satisfactory precision. Meanwhile, using the AI models precise predictions are provided without applying complicated thermodynamic principles.

Research on HC-LSSVM Model for Soft Soil Settlement Prediction Based on Homotopy Continuation Method

Prediction of soft soil settlement is an important research topic in the field of civil engineering, and the least square support vector machine is one of the commonly used prediction methods at present. Nonetheless, the existing LSSVM models have problems of low search efficiency in the search process and lack of global optimal solution in the search results. In order to solve this problem, based on the leave-one-out cross-validation method, the homotopy continuation method was used to optimize the LSSVM model parameters, and then the HC-LSSVM model was constructed with the goal of minimizing the sum of squares of the prediction error of the full sample retention one. Finally, the rationality and correctness of the model are verified by engineering application. The results show that the HC-LSSVM model constructed in this study can accurately predict the settlement of soft ground, which is superior to the common LSSVM model and solves the problem that the parameters of LSSVM model cannot be solved optimally. The research results provide a new method for prediction of soft soil settlement.

A novel grey seasonal model based on cycle accumulation generation for forecasting energy consumption in China

For time series forecasting, the mobile holiday effect typically brings certain difficulty in obtaining accurate forecasts for monthly and quarterly series since it can result in enormous disturbances for modeling, especially for the sequences with limited information and much uncertainty. This study proposes a discrete grey seasonal model by cycle accumulation generation as an alternative approach to seasonal time series forecasting to effectively address such issues. Moreover, the logarithmic transformation technique is introduced to enhance the forecasting performance of this proposed approach, which can significantly reduce the errors resulted from the cycle accumulation generation restoration. To validate the effectiveness and practicality of the proposed model, a range of competing models, including the DGSM(1,1), SFGM(1,1), SGM(1,1), SARIMA, BP neural network, LSSVM, RBF, and LSTM models, are employed to conduct experimental comparisons. Empirical evidence from the monthly industrial electricity and quarterly natural gas consumptions show that the newly designed approach is superior to other competitors in terms of level accuracy and performance stability with the support of the logarithmic transformation technique. Further, the robustness test over the performance of the logarithmic transformation and cycle accumulation generation techniques has been implemented. Results demonstrate that the proposed approach with the logarithmic transformation technique's support can be considered a promising tool to weaken the mobile holiday’ disturbance and improve the forecasting accuracy. Therefore, based on the accurate predictions, the power supply department is able to prepare the power supply plans in advance, reducing the negative effects of the holiday effect on energy management.