NEXT-HOUR ELECTRICITY PRICE FORECASTING USING LEAST SQUARES SUPPORT VECTOR MACHINE AND GENETIC ALGORITHM

It is well known that the performance of Least Squares Support Vector Machine (LSSVM) is guaranteed by k-fold cross-validation (k-CS) or other optimized approaches to choose an appropriate setting of a number of parameters including such as the regularization parameter (7) and the kernel parameter (σ) and so on. However, it is mentioned in a large number of research on k-CS that it need more computational time and large computational burden in the process of optimizing parameters for LSSVM. Moreover, the obtained parameters by CS method are fixed and it lead easily to a poor generalization capabilities in various application. In order to avoid k-CS and to implement parameters update online, this paper proposes a novel method which applied Unscented Kalman Filter (UKF) to dynamically implement parameter updating problem for LSSVM. To estimate LSSVM's parameters online, the state and measurement equations of UKF are first constructed by considering LSSVM's parameter choice as state variable and treating LSSVM model as the measurement equation, respectively. Then, the UKF approach is used to update the LSSVM parameters timely according to the last obtained instance. Applying the proposed method, LSSVM parameters are not any more fixed as tuned parameters on the training dataset, but are adjusted dynamically as new measurements arrived. Finally, the viability and superiority of the proposed method are verified by simulation.

Operations of water distribution networks (WDNs) are monitored daily via installed data loggers, where the collated hydraulic data can be leveraged to improve the system’s operations over time, and to minimize total economic losses due to non-revenue water (NRW). In collaboration with Public Utility Board (PUB), Singapore’s National Water Agency, a practically novel model calibration approach using 24/7 monitoring flow and pressure data has been developed to facilitate PUB’s Smart Water Grid (SWG). The approach is developed as a generic integrated solution process to conduct a series of systematic analyses for daily WDN model calibration, namely: (1) estimating the system’s daily NRW contributions; (2) performing flow calibration that involves net demand consumption calibration, adjusting pumps operational configurations and localizing NRW sources when the system’s daily estimated NRW volume exceeds its assumed background volume; (3) performing energy calibration by rectifying possible drifting in monitored pressure head data and calibrating other physical properties which include, but not limited to, pipe roughness and valve settings, especially during peak-demand hours. The effectiveness of our proposed approach is subsequently tested on three WDN zones in Singapore, having a total pipe length of >100km, that comprises of atypical water usage patterns. The results of model calibration for one of three zones is presented in this paper. The key outcomes derived from the study are: (a) localized a reported leakage event by PUB to less than 100m; (b) calibrated the system’s flow balance, to less than 1% average mean absolute percentage error (MAPE), by first identifying and addressing the system’s billing data uncertainties, followed by localizing anomaly events that account for the total NRW volume estimated; and (c) calibrated the system’s pipe roughness values to improve the total energy balance by achieving an average daily MAPE of 4.0%.

Accurate electricity price prediction is key to the orderly operation of the electricity market. However, the uncertain, stochastic and fluctuant characteristics of electricity pricees make prediction difficult. With the aim of solving this issue, this investigation proposed a multi-stage intelligent model integrating the Beveridge–Nelson decomposition (B-N-D) model, the least square support vector machine (LSSVM), and a nature-inspired optimization model named the whale optimization algorithm (WOA). Firstly, the B-N-D model was utilized to decompose the hourly electricity price time series into determinacy component, periodic trend, and stochastic item. Secondly, the WOA–LSSVM model was proposed to forecast the future hourly data of three components respectively, of which the optimal parameters of LSSVM were determined by using WOA. Finally, the future hourly electricity price data were computed by multiplying the forecasted data of those terms. To verify the validity of the proposed electricity price prediction model in this paper, five comparison approaches based on the B-N-D approach were selected, which are auto-regressive integrated moving average (ARIMA), single LSSVM, LSSVM optimized by the fruit-fly optimization algorithm (FOA), LSSVM optimized by particle swarm optimization (PSO) models, and WOA–LSSVM without B-N-D. By comparatively analyzing the error criteria values of the above models through testing on the objective data of the Pennsylvania–New Jersey–Maryland (PJM) electricity market collected from 11 December 2017 to 18 December 2017, from 15 January 2018 to 22 January 2018, and from 1 February 2018 to 25 February 2018, we conclude that the constructed intelligent model in this paper can greatly enhance the prediction precision of electricity prices.

The application of support vector machine to the aircraft power supply system breakdown diagnosis is one of the research focuses in the tomorrow. Support vector machine (SVM) belongs a new type of machine learning technique, which uses structural hazard minimization rule to substitute for the conventional empirical hazard minimization according to great specimen. The powerful performance of the least squares support vector machine (LSSVM) needs to be reflected in the optimal selection of appropriate parameters, and the quality of parameters greatly affects the accuracy and efficiency of breakdown diagnosis. The traditional LSSVM parameters selection method is inefficient, the calculation is huge, and it takes expensive time to select the satisfactory solution. Aiming at the problem of parameters selection of LSSVM, the way of optimizing the parameters of LSSVM by apply simulated annealing genetic algorithm (SAGA) is proposed. SAGA uses the parallel sampling process of genetic algorithm to improve the time performance of optimization, and makes use of the simulated annealing algorithm to dominate the constriction of majorization to prevent precocity. On the one hand, genetic algorithm has strong ability to grasp the overall search process, on the other hand, simulated annealing algorithm is used to control the convergence of the algorithm to prevent premature phenomenon. The automatic optimal selection of LSSVM parameters is achieved. The Iris system classification and recognition in the UCI machine learning database and the breakdown diagnosis of the autopilot flight control box are used as experimental platforms to obtain data samples for simulation research. The simulation results show that LSSVM optimized by SAGA improves both the accuracy and efficiency of classification recognition. It not only effectively overcomes the problem of low efficiency caused by searching optimal parameters by experience, but also effectively improves the accuracy of classification recognition. The false alarm rate and redundant breakdowns are effectively reduced.

With the in-depth implementation of China’s “National Strategy for Building a Strong Transportation Network,” the scale of expressway construction has continued to expand. As a result, the number of high-fill and deep-cut subgrade projects under complex geological conditions has increased significantly, leading to a surge in landslide-related issues. Consequently, accurate prediction of slope displacement is of critical importance for early warning and prevention of landslide disasters. This study proposes a hybrid prediction model, VMD-MPA-LSSVM-LSTM (VMLL), which integrates Variational Mode Decomposition (VMD), Marine Predators Algorithm (MPA), Least Squares Support Vector Machine (LSSVM), and Long Short-Term Memory (LSTM) networks. Using monitoring data from the high-fill embankment slope at Hongtuyao as the research subject, the VMLL model is employed to predict slope displacement based on small-sample data. The objective is to provide a more accurate method for early warning of slope displacement. Firstly, the original monitoring data are decomposed into trend displacement components and fluctuation displacement components using Variational Mode Decomposition (VMD). Subsequently, the trend component and the fluctuation component are predicted using Least Squares Support Vector Machine (LSSVM) and Long Short-Term Memory (LSTM) networks, respectively. Finally, the Marine Predators Algorithm (MPA) is employed to optimize the hyperparameters of the predictive models. Based on this framework, a VMLL-based slope displacement prediction model is constructed. To verify the superiority of the VMLL model, a comparative analysis was conducted against LSSVM, LSTM, and the VMD-LSSVM-LSTM models. The results demonstrate that the VMLL model achieves the highest prediction accuracy, with a Mean Absolute Percentage Error (MAPE) of 0.4022%, a Mean Absolute Error (MAE) of 0.016 mm, a Root Mean Square Error (RMSE) of 0.0206 mm, a coefficient of determination (R²) of 94.08%, and a Variance Accounted For (VAF) of 96.5%. Compared with the other three models, the VMLL model reduces the MAPE, MAE, and RMSE by 30.41–67.62%, 30.38–67.40%, and 27.32–71.00%, respectively. Meanwhile, it improves the R² and VAF by 5.95–217.51% and 0.38–11.48%, respectively. These improvements clearly demonstrate that the VMLL model outperforms the other models, indicating its significant advantage over single prediction models. Furthermore, three additional datasets were used to evaluate the model’s performance. The average values of the performance evaluation metrics for these datasets were 0.6207% (MAPE), 0.0238 mm(MAE), 0.0273 mm(RMSE), 90.33%(R²), and 96.81%(VAF), respectively. These results demonstrate the high accuracy and strong robustness of the proposed model in predicting both horizontal displacement and vertical settlement of slopes, providing a reliable methodological framework for slope stability assessment and landslide disaster early warning.

As the relationship between geotechnical slope stability and influencing factors is complex and nonlinear, the least squares support vector machines (LS-SVM) is used to establish the nonlinear relation between slope stability and influencing factors. And in consideration of that parameters selection of LS-SVM exerts a major influence on modeling results, the parameters of LS-SVM are optimized by chaos genetic algorithm (CGA). Thus the CGA LSSVM is proposed for forecasting slope stability. Through the comparison between the method and the simple genetic algorithm (SGA) parameters optimization. The result shows that parameters optimization of the CGA has better faster convergence speed, higher prediction precision. And the model is applied to predict the safety factor of the actual slope engineering and the results are well consistent with the actual situation. It is shown that the model is reasonable and feasible.

Time series analysis is an important and complex problem in machine learning. In this paper, least squares support vector machine (LS-SVM) combined with particle swarm optimization (PSO) is used to time series prediction. The LS-SVM can overcome some shortcoming in the multilayer perceptron (MLP) and the PSO is used to tune the LS-SVM parameters automatically. A benchmark problem, Hénon map time series, has been used as an example for demonstration. It is showed this approach can escape from the blindness of man-made choice of the LS-SVM parameters. It enhances the efficiency and the capability of prediction.

High-quality internal experimental materials data are often “small data.” Using external datasets, e.g., data from the literature and other groups, to expand the “small data” to relatively “big data” is particularly important for improving the prediction accuracy of machine learning models. However, the most critical issue is how to use a small amount of internal reliable data to filter external data extracted from multiple sources and obtain optimal datasets with a distribution similar to that of internal data. This is called the multi-source materials data problem. This question was addressed by designing an active learning-based data screening (ALDS) model that is suitable for small material samples. This study used negative expansion materials as subjects. The results show that ALDS can be used to screen the external multi-source dataset and far exceeds traditional outlier filtering methods. The average mean absolute percentage error (MAPE) of the predictive key target property of the negative thermal expansion coefficient (NTEC) was reduced from 4.301 to 0.056. Furthermore, reverse design experiments were conducted on anti-perovskite manganese nitride (AMN), a type of negative thermal expansion material, to prove that the ALDS model can guide the reverse design on a small AMN dataset using the MAPE between ALDS prediction and the ground truth of samples’ three property indicators. This achieves high confidence levels with values of 0.203,0.126, and 0.115. This verifies that the ALDS proposed method improves the effect of ”materials small internal data” in guiding material design and property prediction.

This paper presents a hybrid time-series and adaptive wavelet neural network (AWNN) model for the day-ahead electricity market clearing price forecast. Instead of using price series, one-period continuously compounded return series is used to achieve more attractive statistical properties. The autoregressive moving average with exogenous variables (ARMAX) model is used to catch the linear relationship between price return series and explanatory variable load series, the generalized autoregressive conditional heteroscedastic (GARCH) model is used to unveil the heteroscedastic character of residuals, and AWNN is used to present the nonlinear, nonstationary impact of load series on electricity prices. The Monte Carlo method is adopted to generate more evenly distributed random numbers used for time series and AWNN models to accelerate the convergence. Several criteria such as average mean absolute percentage error (AMAPE) and the variance of forecast errors are used to assess the model and measure the forecasting accuracy. Illustrative price forecasting examples of the PJM market are presented to show the efficiency of the proposed method.

In this paper, an efficient method for the day-ahead electricity price forecasting (EPF) is proposed based on a long-short term memory (LSTM) recurrent neural network model. LSTM network has been widely used in various applications such as natural language processing and time series analysis. It is capable of learning features and long term dependencies of the historical information on the current predictions for sequential data. We propose to use LSTM model to forecast the day-ahead electricity price for Australian market at Victoria (VIC) region and Singapore market. Instead of using only historical prices as inputs to the model, we also consider exogenous variables, such as holidays, day of the week, hour of the day, weather conditions, oil prices and historical price/demand, etc. The output is the electricity price for the next hour. The future 24 hours of prices are forecasted in a recursive manner. The mean absolute percentage error (MAPE) of four weeks for each season in VIC and Singapore markets are examined. The effectiveness of the proposed method is verified using real market data from both markets. The result shows that the LSTM network outperforms four popular forecasting methods and provides up to 47.3% improvement in the average daily MAPE for the VIC market.

Prediction of future trends in financial time-series data are very important for decision making in the share market. Usually financial time-series data are non-linear, volatile and subject to many other factors like local or global issues, causes a difficult task to predict them consistently and efficiently. This paper present an improved Dynamic Recurrent FLANN (DRFLANN) based adaptive model for forecasting the stock Indices of Indian stock market. The proposed DRFLANN based model employs the least mean square (LMS) algorithm to train the weights of the networks. The Mean Absolute Percentage Error (MAPE), the Average Mean Absolute Percentage Error (AMAPE), the variance of forecast errors (VFE) is used for determining the accuracy of the model. To improve further the forecasting results, we have introduced three technical indicators named Relative Strength Indicator (RSI), Price Volume Change Indicator (PVCI), and Moving Average Volume Indicator (MAVI). The reason of choosing these three indicators is that they are focused on important attributes of price, volume, and combination of both price and volume of stock data. The results show the potential of the model as a tool for making stock price prediction.

We propose an efficient framework for optimizing the design of Carbon Nanotube Field-Effect Transistor (CNTFET) through the integration of device physics, machine learning (ML), and multi-objective optimization (MOO). Firstly, we leverage the calibrated TCAD model based on experimental data to dissect the physical mechanisms of CNTFET, gaining insights into its operational principles and unique physical properties. This model also serves as a foundation, enabling multi-scale performance evaluations essential for dataset construction. In the ML phase, a chain structure of Support Vector Regression (SVR Chain) guided by a comprehensive statistical analysis of the design metrics is utilized to predict the design metrics. The surrogate model based on the SVR Chain achieves an average mean absolute percentage error (MAPE) of 1.59% across all design metrics without overfitting, even with limited data. The established ML model exhibits its competence in rapidly producing a global response surface for multi-scale CNTFET. Remarkably, an anomalous equivalent oxide thickness (EOT) and ON-state current (I on ) relationship is observed in CNTFET behavior due to extreme gate length scaling in long channel devices. This intriguing observation is further elucidated through a physics-based explanation. We further compare shallow and deep learning-based TCAD digital twins for model selection guidance. Using the Non-Dominated Sorted Genetic Algorithm-II (NSGA-II) in MOO, we harmonize metrics at both device and circuit levels, significantly reducing the design space. The closed-loop framework expedites the early-stage development of advanced transistors, overcoming the challenges posed by limited data.

This study delves into the environmental impact of inland aquaculture on estuarine ecosystems by examining the water quality of four estuarine streams within the key inland aquaculture zone of South India. In this region, extensive and intensive aquaculture practices are common, posing potential challenges to estuarine health. The research explores the predictive capabilities of the Gaussian elimination method (GEM) and machine learning techniques, specifically multi-linear regression (MLR) and support vector regressor (SVR), in forecasting the water quality index of these streams. Through comprehensive evaluation using performance metrics such as coefficient of determination (R2) and average mean absolute percentage error (MAPE), MLR and SVR demonstrate higher prediction efficiency. Notably, employing key water parameters as inputs in machine learning models is also more effective. Biochemical oxygen demand (BOD) emerges as a critical water parameter, identified by both MLR and SVR, exhibiting high specificity in predicting water quality. This suggests that MLR and SVR, incorporating key water parameters, should be prioritized for future water quality monitoring in intensive aquaculture zones, facilitating timely warnings and interventions to safeguard water quality.

A novel decomposition-ensemble forecasting system for dynamic dispatching of smart grid with sub-model selection and intelligent optimization

This study investigated the use of a Vision Transformer (ViT) for reconstructing GABA-edited Magnetic Resonance Spectroscopy (MRS) data from a reduced number of transients. Transients refer to the samples collected during an MRS acquisition by repeating the experiment to generate a signal of sufficient quality. Specifically, 80 transients were used instead of the typical 320 transients, aiming to reduce scan time. The 80 transients were pre-processed and converted into a spectrogram image representation using the Short-Time Fourier Transform (STFT). A pre-trained ViT, named Spectro-ViT, was fine-tuned and then tested using in-vivo GABA-edited MEGA-PRESS data. Its performance was compared against other pipelines in the literature using quantitative quality metrics and estimated metabolite concentration values, with the typical 320-transient scans serving as the reference for comparison. The Spectro-ViT model exhibited the best overall quality metrics among all other pipelines against which it was compared. The metabolite concentrations from Spectro-ViT's reconstructions for GABA+ achieved the best average R2 value of 0.67 and the best average Mean Absolute Percentage Error (MAPE) value of 9.68%, with no significant statistical differences found compared to the 320-transient reference. The code to reproduce this research is available at https://github.com/MICLab-Unicamp/Spectro-ViT