Energy Data Forecasting Based on the STRLM Time Series Prediction Model

An enterprise service-level performance time series is a sequence of data points that quantify demand, throughput, average order-delivery time, quality of service, or end-to-end cost. Analytical and predictive models of such time series can be embedded into an enterprise information system (EIS) in order to provide meaningful insights into potential business problems and generate guidance for appropriate solutions. Time-series analysis includes periodicity detection, decomposition, and correlation analysis. Time-series prediction can be modeled as a regression problem to forecast a sequence of future time-series datapoints based on the given time series. The state-of-the-art (baseline) methods employed in time-series prediction generally apply advanced machine-learning algorithms. In this article, we propose a new univariate method for dealing with midterm time-series prediction. The proposed method first analyzes the hierarchical periodic structure in one time series and decomposes it into trend, season, and noise components. By discarding the noise component, the proposed method only focuses on predicting repetitive season and smoothed trend components. As a result, this method significantly improves upon the performance of baseline methods in midterm time-series prediction. Moreover, we propose a new multivariate method for dealing with short-term time-series prediction. The proposed method utilizes cross-correlation information derived from multiple time series. The amount of data taken from each time series for training the regression model is determined by results from hierarchical cross-correlation analysis. Such a data-filtering strategy leads to improved algorithm efficiency and prediction accuracy. By combining statistical methods with advanced machine-learning algorithms, we have achieved a significantly superior performance in both short-term and midterm time-series predictions compared to state-of-the-art (baseline) methods.

Satellite Image Time Series (SITS) have recently been of great interest due to the emerging remote sensing capabilities for Earth observation. Trend and seasonal components are two crucial elements of SITS. In this paper, a novel framework of SITS decomposition based on Ensemble Empirical Mode Decomposition (EEMD) is proposed. EEMD is achieved by sifting an ensemble of adaptive orthogonal components called Intrinsic Mode Functions (IMFs). EEMD is noise-assisted and overcomes the drawback of mode mixing in conventional Empirical Mode Decomposition (EMD). Inspired by these advantages, the aim of this work is to employ EEMD to decompose SITS into IMFs and to choose relevant IMFs for the separation of seasonal and trend components. In a series of simulations, IMFs extracted by EEMD achieved a clear representation with physical meaning. The experimental results of 16-day compositions of Moderate Resolution Imaging Spectroradiometer (MODIS), Normalized Difference Vegetation Index (NDVI), and Global Environment Monitoring Index (GEMI) time series with disturbance illustrated the effectiveness and stability of the proposed approach to monitoring tasks, such as applications for the detection of abrupt changes.

Graph representation learning (GRL) has emerged as a unifying paradigm for modeling the relational and heterogeneous nature of battery energy storage systems (BESS), yet a systematic synthesis focused on data-scarce (few-shot) battery scenarios is still lacking. Graph representation learning offers a natural way to describe the structure and interaction of battery cells, modules and packs. At the same time, battery applications often suffer from very limited labeled data, especially for new chemistries, extreme operating conditions and second-life use. This review analyzes how graph representation learning can be combined with few-shot learning to support key battery management tasks under such data-scarce conditions. We first introduce the basic ideas of graph representation learning, including models based on neighborhood aggregation, contrastive learning, autoencoders and transfer learning, and discuss typical data, model and algorithm challenges in few-shot scenarios. We then connect these methods to battery state estimation problems, covering state of charge, state of health, remaining useful life and capacity. Particular attention is given to approaches that use graph neural models, meta-learning, semi-supervised and self-supervised learning, Bayesian deep networks, and federated learning to extract transferable features from early-cycle data, partial charge–discharge curves and large unlabeled field datasets. Reported studies show that, with only a small fraction of labeled samples or a few initial cycles, these methods can achieve state and life prediction errors that are comparable to or better than conventional models trained on full datasets, while also improving robustness and, in some cases, providing uncertainty estimates. Based on this evidence, we summarize the main technical routes for few-shot battery scenarios and identify open problems in data preparation, cross-domain generalization, uncertainty quantification and deployment on real battery management systems. The review concludes with a research outlook, highlighting the need for pack-level graph models, physics-guided and probabilistic learning, and unified benchmarks to advance reliable graph-based few-shot methods for next-generation intelligent battery management.

Simple SummaryForecasting rabbit house environmental variables is critical to achieving intensive rabbit breeding and rabbit house environmental regulation. As a result, this paper proposes a decomposition-based multi-step forecasting model for rabbit houses using a time series decomposition algorithm and a deep learning combinatorial model. The experimental results demonstrated that the proposed method could provide accurate decisions for rabbit house environmental regulation.To improve prediction accuracy and provide sufficient time to control decision-making, a decomposition-based multi-step forecasting model for rabbit house environmental variables is proposed. Traditional forecasting methods for rabbit house environmental parameters perform poorly because the coupling relationship between sequences is ignored. Using the STL algorithm, the proposed model first decomposes the non-stationary time series into trend, seasonal, and residual components and then predicts separately based on the characteristics of each component. LSTM and Informer are used to predict the trend and residual components, respectively. The aforementioned two predicted values are added together with the seasonal component to obtain the final predicted value. The most important environmental variables in a rabbit house are temperature, humidity, and carbon dioxide concentration. The experimental results show that the encoder and decoder input sequence lengths in the Informer model have a significant impact on the model’s performance. The rabbit house environment’s multivariate correlation time series can be effectively predicted in a multi-input and single-output mode. The temperature and humidity prediction improved significantly, but the carbon dioxide concentration did not. Because of the effective extraction of the coupling relationship among the correlated time series, the proposed model can perfectly perform multivariate multi-step prediction of non-stationary time series.

Prediction of anomaly in time-varying datasets is at the core of many real-world applications. For example, prediction of anomalous events in different time-varying business metrics and proactively mitigating any underlying problem in the system is of utmost importance to any online business to evade any kind of disruption in smooth customer experience. A large body of literature exists which adopted different techniques for anomaly prediction in multi-variate time-series data. Recent advancements in deep-learning have resulted in multiple literature proposing different variants of graph neural network (GNN) for anomaly prediction in multi-variate time-series data. However, one major gap area that remained in this domain is the sensitivity of such networks to inherent noise in the data. This is where we make our contribution by proposing a graph neural network with dynamic attention which not only offers more robustness of the network to noise but also makes the network more trainable and accurate in prediction in the presence of noise. We carried out extensive experimentation against both publicly available datasets and real-world industrial datasets. The outcome of the experimentation is consistently in support of the superiority of our proposed network over the state-of-the-art GNN network for anomaly prediction.

The increasing penetration of behind-the-meter (BTM) photovoltaic (PV) systems introduces substantial uncertainty into power networks. This study proposes a time series decomposition-based hybrid model to accurately estimate the gross load in environments where BTM PV data are unavailable. The proposed model first performs time series decomposition to separate the input sequence into trend and seasonal components. The trend component is fed into a bidirectional long short-term memory (BiLSTM), while the seasonal component is processed through a Transformer to capture complex periodic patterns. The outputs of the two pathways are fused through a cross-gate mechanism, and the combined representation is further refined by a residual BiLSTM for nonlinear residual correction. With its simple structure, the proposed approach provides a practical solution for gross load estimation in environments without BTM PV measurements.

Time series analysis and prediction is an important means of dynamic system modeling. A new method of time series prediction based on support vector regression (SVR) is introduced to resolve the problem of non-linear system modeling. For the purpose of reducing calculation complexity, smooth method is presented to improve standard SVR arithmetic, and is utilized to build the combustion state model of flame in the furnace of utility boilers according to the feature parameters of flame image, in order to predict the combustion state of flame. The flame images are gained from the flame image gathering system on-line. The feature parameters of the flame image are extracted, and are used to determined combustion indices which can denote different combustion states of flame. The time series of combustion indices are used for constructing the smooth support vector regression (SSVR) model and predicting the combustion state of flame. The results of experimentation indicate that SSVR has excellent performance on time series prediction. Compared with traditional time series prediction method such as artificial neural network, SSVR has faster convergence speed and higher fitting precision, which effectively extends the application of SVR.

Traditional Graph Neural Network (GNN) learning patterns can only achieve optimal performance under the assumption of independent and identically distributed data. However, in real-world scenarios, numerous unpredictable factors give rise to the Out-of-Distribution (OOD) problem. Moreover, this distribution discrepancy may lead to unreliable predictions by GNNs models within unknown OOD domains. Therefore, this paper proposes a machine learning method to enhance the stability of GNN cross-site classification from the perspective of stable learning. The aim is to extract genuine causal features while eliminating spurious causal features. By introducing a feature sample weighting decorrelation technique in the random Fourier transform space and combining it with a baseline GNN model, a Stable-GNN model and a constrained sampling weight gradient update algorithm are designed. Its theoretical proof indicates that this algorithm can ensure the decrease in the loss, thus remedying the shortcomings of the original algorithm in weight updating. While guaranteeing predictive performance on the training distribution data, it also reduces prediction bias on data from unseen test distributions. Compared with the GNN model and the Stable-GNN model (S-GNN), the experimental results of the S-GNN model demonstrate that it not only surpasses the current state-of-the-art GNN models, but also offers a flexible framework for strengthening existing GNNs.

A time series prediction scheme based on multiresolution-based bilinear recurrent neural network (MBLRNN) is proposed in this paper. The proposed predictor is based on the BLRNN that has been proven to have robust abilities in modeling and predicting time series. The learning process is further improved by using a multiresolution-based learning algorithm for training the BLRNN so as to make it more robust for long-term prediction of the time series. The proposed MBLRNN-based predictor is applied to the long-term prediction of time series. Experiments and results on the Mackey-Glass Series data and Sunspot Series data show that the proposed MBLRNN outperforms both the traditional multilayer perceptron type neural network (MLPNN) and the BLRNN in terms of the normalized mean square error (NMSE).

Many German cities, municipalities and transport associations are expanding their bike-sharing systems (BSS) to offer citizens a cost-effective and climate-friendly means of transport and an alternative to private motorized transport (PMT). However, operators face the challenge of generating high-quality predictive analyses and time series forecasts. In particular, the prediction of demand is a key component to foster data-driven decisions. To address this problem, an Unobserved Component Model (UCM) has been developed to predict the monthly rentals of a BSS, whereby the station-based BSS VRNnextbike, including over 2000 bikes, 297 stations and 21 municipalities, is employed as an example. The model decomposes the time series into trend, seasonal, cyclical, auto-regressive and irregular components for statistical modeling. Additionally, the model includes exogenous factors such as weather, user behavior (e.g., traveled distance), school holidays and COVID-19 relevant covariates as independent effects to calculate scenario based forecasts. It can be shown that the UCM calculates reasonably accurate forecasts and outperforms classical time series models such as ARIMA(X) or SARIMA(X). Improvements were observed in model quality in terms of AIC/BIC (2.5% to 22%) and a reduction in error metrics from 15% to 45% depending on the considered model.

This paper presents techniques for studying the influence of the climatic and other variables for the explanation of the water use with an example of time series in Gainesville, Florida. A statistical methodology is described for separating the different time scale components in time series of water use, namely, long term component, seasonal component, and short term component. We analyze each component separately and we prove that the temperature, precipitation, soil temperature, and relative humidity time series are the main climatic factors for the explanation of the long term, seasonal and short term component of the water use time series. Part of the residuals derived from the linear regression of the long term component of the water use can be explained by the unemployment rate. We also show that with the decomposition of the water use time series the explanation of the water use has been improved approximately two times. The explanation of the long term component of water use by the long term regional weather parameters can enable us to the long term regional prediction of the water resources availabilities. This methodology can be applied for studying the water use time series in other locations, as well.

The graph neural network (GNN) has recently become an emerging research topic for processing non-euclidean data structures since the data used in various popular application domains are usually modeled as a graph, such as social networks, recommendation systems, and computer vision. Previous GNN accelerators commonly utilize the hybrid architecture to resolve the issue of “hybrid computing pattern” in GNN training. Nevertheless, the hybrid architecture suffers from poor utilization of hardware resources mainly due to the dynamic workloads between different phases in GNN. To address these issues, existing GNN accelerators adopt a unified structure with numerous processing elements and high bandwidth memory. However, the large amount of data movement between the processor and memory could heavily downgrade the performance of such accelerators in real-world graphs. As a result, the processing-in-memory architecture, such as the ReRAM-based crossbar, becomes a promising solution to reduce the memory overhead of GNN training. In this work, we present the TCAM-GNN, a novel TCAM-based data processing strategy, to enable high-throughput and energy-efficient GNN training over ReRAM-based crossbar architecture. Several hardware co-designed data structures and placement methods are proposed to fully exploit the parallelism in GNN during training. In addition, we propose a dynamic fixed-point formatting approach to resolve the precision issue. An adaptive data reusing policy is also proposed to enhance the data locality of graph features by the bootstrapping batch sampling approach. Overall, TCAM-GNN could enhance computing performance by 4.25× and energy efficiency by 9.11× on average compared to the neural network accelerators.

Prediction of the time series from short data samples was considered. An expert-statistical approach based on specially designed methods for joint use of the subjective and objective data was proposed for handling this problem. Its solution relies on the scheme of expert-statistical processing called the analog method.

Extensive studies have been conducted to investigate Artificial Intelligence (AI) in the context of time series data. In this article, we investigate the complex domain of industrial time series, from the dimensions of classical machine learning (ML), deep neural networks (DNNs) and graph neural networks (GNNs). Current surveys often focus on a specific methodology or oversee the connection of diverse approaches; our article bridges this gap by providing an all-inclusive interpretation across numerous techniques. In addition, the aim of this article is to focus on the core areas of time series such as forecasting, classification, and anomaly detection. From traditional methodologies like Autoregressive Integrated Moving Average (ARIMA) and support vector machine (SVM) methods, the advancements of DNNs, for instance long-short-term memory (LSTMs), convolutional neural networks (CNNs), attention mechanisms, and transformers, describe how temporal information is used for forecasting, anomaly detection, and classification. Then the article discusses the advances and limitations in ML, DNN, and GNN in order to improve the different methods in either category. Lastly, we outline future directions and open research questions with the different methodologies used in time series.

Marine monitoring faces unprecedented challenges as climate change and human activities reshape ocean ecosystems. Traditional tracking methods struggle with the scale and complexity of modern marine sensing needs. This paper proposes distributed networks of low-cost drifting sensors and presents a comparative study of heterogeneous graph neural networks (GNNs) versus Kalman filters for predicting marine entity trajectories in such distributed networks with intermittent observations. Using the Aquascan simulation framework, we model drifting sensors detecting marine entities across 480 km² of ocean surface with realistic movement patterns and sparse coverage. Experiments across multiple prediction horizons show GNNs significantly outperform Kalman filters: GNNs maintain over 95% AUC (Area Under the Curve) across all horizons while Kalman filters degrade from 97% to 69% AUC. The performance gap widens under challenging conditions—at 5 km sensor spacing, GNNs achieve 92.8% AUC versus 66.9% for Kalman filters. GNNs' superior performance stems from leveraging network topology and reasoning about non-detections to infer entity presence in coverage gaps. These results demonstrate that graph-based approaches offer substantial advantages for distributed marine monitoring.