Singular Value Decomposition Technique Research Articles

Energy storage battery state of health estimation based on singular value decomposition for noise reduction and improved LSTM neural network

Accurate estimation of the State of Health (SOH) of batteries is important for intelligent battery management in energy storage systems. To solve the problems of poor quality of data features as well as the difficulty of model parameter adjustment, this study proposes a method for estimating the SOH of lithium batteries based on denoising battery health features and an improved Long Short-Term Memory (LSTM) neural network. First, in this study, three health features related to SOH decrease were selected from the battery charge/discharge data, and the singular value decomposition technique was applied to the noise reduction of the features to improve their correlation with the SOH. Then, the whale optimization algorithm is improved using cubic chaotic mapping to enhance its global optimization-seeking capability. Then, the Improved Whale Optimization Algorithm (IWOA) is used to optimize the model parameters of LSTM, and the IWOA-LSTM model is applied to the battery SOH estimation. Finally, the model proposed in this research is validated against the Center for Advanced Life Cycle Engineering (CALCE) battery dataset. The experimental results show that the prediction error of battery SOH by the method proposed in this study is less than 0.96%, and the prediction error is reduced by 49.42% compared to its baseline model. The method presented in the article achieves accurate estimation of the SOH, providing a reference for practical engineering applications.

Accurate and efficient daily carbon emission forecasting based on improved ARIMA

Major nations across the globe are increasingly concerned about the rising trends in carbon dioxide (CO2) emissions, particularly in societies of varying scales. Against this backdrop, precise prediction of carbon emissions becomes critically important, especially for the formulation and adjustment of near-term carbon reduction policies. However, the non-linear, non-stationary, and complex nature of daily carbon emission data poses a great challenge for daily-level forecasting especially in the big data context. To address this issue, we propose a novel composite forecasting approach named DCEF dedicated to the estimation of daily carbon emissions. In concrete, our approach employs the Empirical Mode Decomposition (EMD) for data stabilization and the Auto-regressive Integrated Moving Average (ARIMA) model for forecasting, while integrating the Truncated singular value decomposition(TSVD) technique for data compression and mitigating noise. Finally, DCEF is empirically validated with real daily carbon emission datasets collected from 6 sectors in 13 countries of varying sizes. Experimental results demonstrate the advantages of our approach compared to other baseline models in terms of prediction accuracy and efficiency.

Mittag-Leffler function based security control for fractional-order complex network system subject to deception attacks via Observer-based AETS and its applications

The goal of this paper is to investigate the security control for uncertain fractional-order delayed complex network systems under deception attacks using the Mittag-Leffler function and observer-based adaptive event-triggered scheme (AETS) with the fractional commensurate order in q ∈ (0, 1). The adaptive event-triggering scheme is used during the data transmission process from the sensors to the observer, where the triggering threshold can be dynamically modified to reduce resource waste. We make a novel model for the estimation error system that takes into account both the effects of the adaptive event-triggered scheme and the effects of deception attacks. A sufficient condition is obtained to guarantee the stochastic mean-square stability of the augmented error system using the Mittag-Leffler (M-L) functions and the Lyapunov functional method and by using the singular value decomposition (SVD) and linear matrix inequality (LMI) techniques, the co-design problem of desired observer and controller gains is found, and it is shown that the solution ensures the stability of a closed-loop uncertain fractional-order complex networked system. At the end of this study, two numerical examples and diesel engine system model are given to show that the above findings are correct.

Transient amplification due to non-Hermitian interference of dissipative supermodes

Achieving transient amplification of light in coupled waveguides that are characterized by a fully dissipative spectrum, is the context of our study. Physically, material gain is present and spatially distributed in a lossy lattice, and thus makes possible the interesting effect of transient growth. In order to quantify the maximum value of power growth and determine the accompanying optimal initial conditions, we apply pseudospectra and singular value decomposition techniques. Similar to the way we can identify all possible microstates of a paramagnetic crystal for a given value of energy, we can also calculate the maximum transient growth of all different permutations of the gainy channels for given values of gain–loss amplitudes, and thus sort them. The result of such non-Hermitian hierarchy is that we have maximum amplification, when the gain is distributed close to the lattice's edges.

Oblique Compressible Waves in the Reconnection Exhaust Region Embedded in the Inner Heliospheric Current Sheet Observed by Parker Solar Probe

Magnetic reconnection is an important physical process of energy conversion in the heliosphere. Parker Solar Probe (PSP) passes through current sheets of the inner heliosphere and is likely to encounter magnetic reconnection events there. PSP traversed a magnetic reconnection exhaust region that occurred in the coronal streamer during its perihelion Encounter 8. We report an observation of the counterstream of strahl electrons and compressible waves in the exhaust region on the antisunward side of the reconnection site. We analyze the wave characteristics using electromagnetic singular value decomposition techniques and find that the propagation direction of the compressible waves is quasi-perpendicular to the local magnetic field. Combining with the topology of the magnetic field, we infer that the compressible waves converge from the edge to the center of the exhaust region, and then propagate away from it. Further, we select 12 magnetic reconnection events during Encounter 5–8 for statistics and find that the oblique compressible waves are commonly detected throughout the inner heliospheric current sheet. In addition, we discuss the possible nature of wave branches for these compressible waves. Our work shows that magnetic reconnection in the heliosphere not only changes the topology of the large-scale magnetic field in the heliosphere, but also affects the transport characteristics of solar wind plasma and suprathermal particles, and regulates the states of waves and turbulence in the heliosphere.

A non-convex low-rank image decomposition model via unsupervised network

Image decomposition is the separation of a given image into two parts with different features, i.e., structure and texture. In order to extract the image information comprehensively, a non-convex unsupervised image decomposition model is proposed in this paper. Firstly, we propose a new non-convex low-rank regular term to describe the low-rank nature of images, which focuses on optimizing the sum of partial singular values of image matrices using a non-convex truncated nuclear norm. Secondly, unsupervised network is used to capture image information by learning the deep semantic information of the image and then extracting the image structure. Finally, the alternative direction method of multipliers is used to divide the model into three sub-problems and optimize them separately. For the non-convex sub-problem, a two-stage solution scheme is given by employing the singular value decomposition technique, and experimental analysis is carried out on both real and synthetic images. The experimental results show that the proposed non-convex image decomposition method outperforms other state-of-the-art methods in terms of subjective visual effects. Meanwhile, in the quantitative evaluation, the proposed method shows a large improvement compared with other methods, with a maximum improvement of 15 % in the structural similarity index measure (SSIM) index.

Structural damage detection based on structural macro-strain mode shapes extracted from non-stationary output responses

Abstract Long-gauge fiber Bragg grating strain sensors have been widely employed because of their broader measuring range and higher sensitivity. However, current structural damage detection methods using macro-strain modal parameters are based on structural frequency response function or stationary power spectrum density, which are not applicable to non-stationary responses. To overcome this limitation, an improved method is proposed in this paper for structural damage detection based on structural macro-strain responses under unknown multi-point non-stationary excitations. First, a new concept of macro-strain energy spectrum transmissibility (MEST) is proposed using structural non-stationary macro-strain responses, and it is derived that MEST at a certain system pole equals the ratio of macro-strain mode shape. Then, the singular value decomposition technique is adopted for the MEST matrix to identify structural natural frequencies and macro-strain mode shapes. Finally, two damage detection indicators are constructed based on the identified normalized macro-strain mode shape (NMMS). The first indicator is the difference in structural NMMS before and after structural damage. The second one is based on the curvatures of structural NMMS, which can be used for structures without intact baseline. Numerical verifications are conducted to identify beam-type structural damage under multi-point non-stationary excitations or vehicle loads. Five damage scenarios with different measurement noise levels are investigated, and damage detection results validate the effectiveness of the proposed method.

Finite-time master–slave synchronization for implicit hybrid neural networks under event-triggered guaranteed cost control and random deception attacks

This paper researches the finite-time master–slave synchronization for implicit hybrid neural networks under event-triggered guaranteed cost control and random deception attacks. By developing Lyapunov–Krasovskii functional and exploiting singular value decomposition technique, the finite-time boundedness of synchronization error closed-loop system is achieved under deception attacks and event-triggered scheme, then the finite-time master–slave synchronization of implicit Markovian jump neural networks is realized. The desired gains of state feedback synchronization guaranteed cost controller are designed by solving a set of linear matrix inequalities. The availability of the given approach is confirmed by an artificial neural network and a numerical example.

Singular Value Decomposition-Based Penalized Multinomial Regression for Classifying Imbalanced Medulloblastoma Subgroups Using Methylation Data.

Medulloblastoma (MB) is a molecularly heterogeneous brain malignancy with large differences in clinical presentation. According to genomic studies, there are at least four distinct molecular subgroups of MB: sonic hedgehog (SHH), wingless/INT (WNT), Group 3, and Group 4. The treatment and outcomes depend on appropriate classification. It is difficult for the classification algorithms to identify these subgroups from an imbalanced MB genomic data set, where the distribution of samples among the MB subgroups may not be equal. To overcome this problem, we used singular value decomposition (SVD) and group lasso techniques to find DNA methylation probe features that maximize the separation between the different imbalanced MB subgroups. We used multinomial regression as a classification method to classify the four different molecular subgroups of MB using the reduced DNA methylation data. Coordinate descent is used to solve our loss function associated with the group lasso, which promotes sparsity. By using SVD, we were able to reduce the 321,174 probe features to just 200 features. Less than 40 features were successfully selected after applying the group lasso, which we then used as predictors for our classification models. Our proposed method achieved an average overall accuracy of 99% based on fivefold cross-validation technique. Our approach produces improved classification performance compared with the state-of-the-art methods for classifying MB molecular subgroups.

Model order reduction, a novel method using krylov sub-spaces and genetic algorithm

Model Order Reduction (MOR) of complex and large systems in Electrical engineering, continuous to be an attractive field for Engineers and Scientists over the last few decades, this complexity of models makes the control designs and simulation using Computer Aided Design (CAD) more and more difficult and consuming a lot of time. There for, accurate, robust and fast algorithms for simulation are needed. The goal of MOR is to replace the original system by an appropriate reduced system which preserves the main properties of the original one such that stability and passivity. Several analytical MOR techniques have been proposed in the literature over the past few decades, to approximate high order linear dynamic systems like Krylov sub-space techniques and SVD (Singular Value Decomposition) techniques. However, most of these techniques lead to computationally demanding, time consuming, iterative procedures that usually result in non-robustly stable models with poor frequency response resemblance to the original high order model in some frequency ranges. Recently a set of new techniques based on Artificial Intelligence (AI) were proposed in [1] for MOR. This article considers the problem of model order reduction of Linear Time In varying (LTI) systems. It is described by first and second order ordinary differential equations model. A tow steps method for model order reduction of LTI systems is proposed here, which combined features of an analytic technique (Krylov approach) and an AI technique (Genetic Algorithm). In the first step, the size of the original model is reduced to an intermediate order, using an analytical technique based on Krylov sub-spaces. In the final step of the reduction process, an AI approach based on Genetic Algorithm (GA) is applied to obtain an optimized nominal model.

Symmetric circulant matrix decomposition-based multivariable group sparse coding for rolling bearing fault diagnosis

Singular value decomposition technique proves its effectiveness in mechanical signal analysis by decomposing the test signal into a series of singular spectral components of different frequency bands. Nevertheless, how to adapt this technology to the needs of cyclo-nonstationary signal and how to set the decomposition number while maintaining detailed features to obtain the optimal component containing the most fault information, remains an important issue that needs to be addressed in the field of mechanical fault diagnosis. To overcome these disadvantages, the symmetric circulant matrix decomposition (SCMD) is presented. Two main ideas structure the present technique. Firstly, symmetric circulant matrix is used to generate eigenvectors, which will better adapt to the cyclo-nonstationary signal associated with the structural symmetry of rotating machinery. Then, an impulse fluctuation measure is established to adaptively search for the decomposition number and extract the optimal component. Moreover, to better improve the impulse extraction effect of SCMD, the multivariate group sparse coding based on the multivariate correlation characteristics and intra group sparsity characteristics of impulse signals is proposed, which can enhance impulse features while preserving fault details as much as possible. The reliability and feasibility of the proposed method are verified by the experimental signals. The comparison with several classic methods shows that this method is more effectiveness in weak feature extraction.

Faster nonconvex low-rank matrix learning for image low-level and high-level vision: A unified framework

This study introduces a unified approach to tackle challenges in both low-level and high-level vision tasks for image processing. The framework integrates faster nonconvex low-rank matrix computations and continuity techniques to yield efficient and high-quality results. In addressing real-world image complexities like noise, variations, and missing data, the framework exploits the intrinsic low-rank structure of the data and incorporates specific residual measurements. The optimization problem for low-rank matrix learning is effectively solved using the nonconvex Proximal Block Coordinate Descent (PBCD) algorithm, resulting in nearly unbiased estimators. Rigorous theoretical analysis ensures both local and global convergence. The PBCD algorithm updates blocks of variables iteratively with closed-form solutions, adeptly handling nonconvexity and promoting faster convergence. Notably, the framework incorporates the randomized singular value decomposition (RSVD) technique and introduces a continuous strategy for adaptive model parameter updates. These strategic choices reduce computational complexity while maintaining result quality. They offer fine-tuned control over the desired rank of the learned matrix and enhance robustness in a straightforward manner. Furthermore, the versatility of the proposed nonconvex PBCD algorithm extends to handling problems with multiple variables, as supported by theoretical analysis. Experimental evaluations, spanning various image low-level and high-level vision tasks such as inpainting, classification, and clustering, validate the effectiveness and efficiency of our framework across diverse databases. The source code is available at https://github.com/ZhangHengMin/FNPBCD_LR.In a nutshell, our framework provides a unified solution to tackle both low-level and high-level vision tasks in images. By combining fast nonconvex low-rank matrix learning with adaptive parameter updates, we achieve efficient computation, yielding high-quality results that demonstrate robustness against various types of noise. The evaluations further endorse the reliability and applicability of our proposed framework.

Full-rank orthonormal bases for conditionally positive definite kernel-based spaces

In kernel-based approximation, it is well-known that the direct approach to interpolation is prone to ill-conditioning of the interpolation matrix. One simple idea is to use other better-conditioned bases that span the same space of the translated kernels i.e. their associated native space. Pazouki and Schaback (2011) tracked this issue by investigating different factorization of the interpolation matrix in order to build stable and orthonormal bases for the corresponding native space of the positive definite kernels. In this paper, we work with the reproducing kernel K for the associated native Hilbert space NΦ(Ω) corresponding to a conditionally positive definite kernel Φ on the nonempty set Ω. We give a well-organized matrix formulation of the kernel matrix K by constructing the matrices corresponding to cardinal basis from monomials. Then, we present two possible ways to find full-rank data-dependent orthonormal bases that are discretely ℓ2 and NΦ-orthonormal. The first approach is given by the factorization of the kernel matrix K and the next one is based on the eigenpairs approximation of the linear operator associated with the reproducing kernel K given by Mercer’s theorem. In the sequel, we employ the truncated singular value decomposition technique to find an optimal low-rank basis with the coefficient matrix whose rank is less than that of the original matrix. Special attention is also given to error analysis, duality, and stability. Some numerical experiments are also provided.

Capacity prediction of Lithium-ion batteries based on adaptive sliding window pooling extreme learning machine algorithm

Abstract Considering the safety and reliability, it is especially important to accurately predict the capacity decline trend of lithium-ion batteries. In this paper, a simple and easy-to-operate singular value decomposition technique is used to extract the health indicators (HIs) that are correlated with the capacity from the measurable parameters of battery, and then the HIs that have a high Pearson correlation coefficient with the capacity are selected for predicting the battery capacity. Aiming at the problems of low prediction accuracy and random dispersion of traditional extreme learning machine (ELM), this paper proposes an adaptive sliding window pooling extreme learning machine (ASW-PELM) algorithm. The algorithm first adaptively adjusts the window length according to the fluctuation of local data, and then dynamically traverses the data with the sliding window for data enhancement, and this adaptive sliding window mechanism provides effective data for the model prediction stage. Then it combines the pooling operation and the ELM to replace the random factor between the input layer and the hidden layer, which effectively solves the problem of random dispersion in the original learning model. The results of lithium battery capacity prediction under two sets of different experimental conditions show that the method has the highest prediction accuracy compared with other generalized algorithms.

Analyzing the Impact of COVID-19 on Liver Cancer: A Comprehensive Study using XGBoost Classifier and Feature Selection Techniques

Healthcare systems around the world have faced challenges due to the Coronavirus disease (COVID-19) epidemic, which has taken resources and attention away from long-term diseases like liver cancer. To identify the effect of COVID-19 on liver cancer, proposed an effective research work with Extreme Gradient Boosting (XGBoost) classifier. The identification of the most suitable feature is executed through the utilization of SelectKBest, Principal Component Analysis (PCA) and Singular Value Decomposition (SVD) techniques. A comparative analysis is carried out to ascertain the more precise feature selection, employing Decision Tree, K-Nearest Neighbors (KNN), CatBoost, LightGBM, and XGBoost algorithms. Further, the proposed work undertakes an analysis of the impact of COVID-19 on liver cancer, encompassing both the pandemic and pre-pandemic periods. In comparison to other feature selection techniques and models, SelectKBest feature selection with Xgboost classifier provided an accuracy of 0.92. The proposed research work provides valuable insights to healthcare professionals and policy makers, providing a better understanding of the challenges faced by liver cancer patients during the pandemic. By meticulously exploring feature selection techniques and employing advanced machine learning models, this study quantifiably demonstrates the effectiveness of the proposed methodology in addressing the impact of COVID-19 on liver cancer, contributing to a more comprehensive understanding of this critical issue.

Artificial Intelligence Empowers Emotional Expression and Aesthetic Imagery in Modern Chinese Literature

Abstract This paper explores how artificial Intelligence enhances emotional expression and aesthetic imagery in modern Chinese literature. The study focuses on the latent semantic analysis of literary texts, text mining challenges, techniques and algorithms for latent semantic analysis, and the application of artificial Intelligence in literary criticism. Through text representation, analysis of text mining challenges, latent semantic analysis (LSA), and the application of specific algorithms, this study provides an in-depth understanding of the intrinsic semantics of literary texts. The study employs text representation methods, such as the TF-IDF formulation, to deal with high-dimensional sparse text data. The main challenges include high dimensionality and discovering potential semantic relationships between words. Latent semantic analysis (LSA) realizes data dimensionality reduction through SVD singular value decomposition technique to extract the implicit relationships between words in text. The results show that LSA effectively pulls the potential semantic structure of the text, reduces the “noise”, simplifies the text vectors, and realizes the dimensionality reduction. In addition, the text clustering analysis using Kohonen self-organized feature mapping network reveals the semantic relationship and sentiment distribution among literary works. Intelligent technology can effectively enhance literary works’ emotional depth and aesthetic value and provide a new interpretation perspective for modern literature.

A Noninvasive Measurement Technique of Grid and Converter Wideband Impedance

A novel technique has been proposed to simultaneously estimate the wideband Thevenin's impedance of the grid as well as the power electronic converter from measurements sampled at the point of common coupling. The proposed algorithm is non-invasive, thus it can estimate both the impedance of the utility grid as well as the grid-tied inverter from the voltage and current measurements at the point of common coupling. It works by utilizing the harmonic information obtained from the existing voltage and current waveforms. Therefore it can circumvent the requirement of any injecting device to estimate the wideband impedance. The algorithm works on an initial <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">Predictor Generation</i> phase, which is an initialization phase based on the design parameters, and a <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">Realtime Estimation</i> phase which runs after the inverter is commissioned. The Realtime Estimation phase uses the Singular Value Decomposition technique to efficiently utilize the existing harmonic information for the prediction. The efficacy of the technique has been tested in EMTDC/PSCAD simulations of a grid-tied inverter system. Furthermore, a Power Hardware in Loop experimental set-up has been built to validate the simulation results and thus, prove the efficacy of the proposed algorithm. The results show a mean square error of 0.05 <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex-math notation="LaTeX">$\Omega$</tex-math></inline-formula> and 0.5 <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex-math notation="LaTeX">$\Omega$</tex-math></inline-formula> for the grid and inverter side respectively.

Research on the Teaching Reform Path of Art and Design Specialties in Colleges and Universities in the Era of Artificial Intelligence

Abstract In this study, the ability level of art and design students in colleges and universities was accurately calculated by adopting the Logistics model in item response theory, combined with learner-question response data, and the maximum likelihood estimation method. Further, the study constructs a personalized recommendation model based on students’ competence, which utilizes the singular value decomposition technique to calculate students’ similarity coefficients, and achieves accurate personalized learning path recommendation. Through descriptive analysis, this paper delves into the specific application and effect of this personalized recommendation algorithm in the teaching reform of art and design majors in colleges and universities. The simulation analysis method evaluates the impact of online art and design professional teaching. The results show that this algorithm performs best in scoring prediction accuracy when the regularization coefficient is set to 0.02, and its RMSE value reaches 0.95012.This finding confirms that the personalized recommendation algorithm based on the students’ ability can effectively promote the reform of art and design teaching in colleges and universities, and realize tailored teaching. In addition, through accurate learning path recommendation, the algorithm significantly improves students’ learning ability and enthusiasm and provides valuable guidance and reference for teaching reform in colleges and universities.

Matrix Factorization For Augmented Deep Learning Model Generation

This paper aims to address two significant challenges of Deep Learning(DL) model generation, high computational cost involved during the training phase and data interpretability from high dimensional data. The computational cost is curbed using centralized as well as distributed matrix factorization technique known as CUR decomposition. CUR decomposition has the advantage of reducing the dimensions of the dataset without data transformation and minimal information loss. Though Singular Value De-composition(SVD) is the best matrix decomposition technique with least reconstruction error, it transforms the data and hence the original data cannot be interpreted from the decomposed matrices while applying DL technique. So CUR decomposition is leveraged to reduce the number of features in high-dimensional data while preserving the essential information. Extensive experimental analysis is conducted to evaluate and compare the performance of CUR and SVD techniques based on reconstruction error and time complexity in centralized as well as distributed setting. The results exhibit that CUR outperforms SVD in terms of computational efficiency, especially in a distributed setting where the decomposition time for CUR was a fraction compared to SVD. At the same time DL models generated from this reduced dataset exhibit comparable accuracy with the original high dimensional dataset as well as reduced dataset obtained through SVD. This work also enhances the accessibility of the CUR algorithm to users, for which an application is developed that enables users to execute the algorithm on any dataset. The user-friendly interface of the application facilitates the upload of datasets, configuration of desired decomposition rank, and execution of the algorithm with a single click.

Event-triggered synchronization for discrete-time delayed neural networks via aperiodic detection

The focus of this paper is to analyze the event-triggered synchronization problem for discrete-time neural networks with time-varying delays. First, an economic event generator is constructed, where the signals are generated by aperiodic detection. The transmission of signals is determined by an aperiodic event-triggered mechanism. This method avoids the continuous sampling and computing for sensors, which results in the saving of computing and communication resources. Second, by taking a piecewise Lyapunov functional, some sufficient conditions are established to ensure the synchronization of the discrete-time neural networks. Compared with the existing results, the proposed Lyapunov functional takes the sawtooth constraint of sampling error signals into account. Third, the control gain and triggering parameter are co-designed by the singular value decomposition technique. Finally, a simulation example is provided to demonstrate the effectiveness of the proposed method.