Search Results (3,591)

Aim: Stock price prediction remains a highly challenging task due to the complex and nonlinear nature of financial time series data. While deep learning (DL) has shown promise in capturing these nonlinear patterns, its effectiveness is often hindered by the low signal-to-noise ratio inherent in market data. This study aims to enhance the stock predictive performance and trading outcomes by integrating Singular Spectrum Analysis (SSA) with deep learning models for stock price forecasting and strategy development on the Australian Securities Exchange (ASX)50 index. Method: The proposed framework begins by applying SSA to decompose raw stock price time series into interpretable components, effectively isolating meaningful trends and eliminating noise. The denoised sequences are then used to train a suite of deep learning architectures, including Convolutional Neural Networks (CNN), Long Short-Term Memory (LSTM), and hybrid CNN-LSTM models. These models are evaluated based on their forecasting accuracy and the profitability of the trading strategies derived from their predictions. Results: Experimental results demonstrated that the SSA-DL framework significantly improved the prediction accuracy and trading performance compared to baseline DL models trained on raw data. The best-performing model, SSA-CNN-LSTM, achieved a Sharpe Ratio of 1.88 and a return on investment (ROI) of 67%, indicating robust risk-adjusted returns and effective exploitation of the underlying market conditions. Conclusions: The integration of Singular Spectrum Analysis with deep learning offers a powerful approach to stock price prediction in noisy financial environments. By denoising input data prior to model training, the SSA-DL framework enhanced signal clarity, improved forecast reliability, and enabled the construction of profitable trading strategies. These findings suggested a strong potential for SSA-based preprocessing in financial time series modeling. Full article

We investigate whether sophisticated deep learning architectures justify their computational cost for short-term cryptocurrency price forecasting. Our study evaluates a 2.1M-parameter (M represents millions (e.g., 2.1M = 2,100,000 parameters), with all RMSE values reported in USD) wavelet-enhanced transformer that decomposes the Fear and Greed Index (FGI) into multiple timescales before integrating these signals with technical indicators. Using Diebold–Mariano tests with HAC-corrected variance, we find that all models—including our wavelet–transformer, ARIMA, XGBoost, LSTM, and vanilla Transformer—fail to significantly outperform the O(1) naive persistence baseline at the 1-day horizon (DM statistic = +19.13, $p<0.001$, naive preferred). Our model achieves an RMSE of USD 2005 versus USD 1986 for naive (ratio 1.010), requiring 3909× more inference time (2.43 ms vs. 0.0006 ms) for a statistically worse performance. These results provide strong empirical support for the Efficient Market Hypothesis in cryptocurrency markets: even sophisticated multi-scale architectures combining wavelet decomposition, cross-attention, and auxiliary technical indicators cannot extract profitable short-term signals. Through systematic ablation, we identify positional encoding as the only critical architectural component—its removal causes 30% RMSE degradation. Our findings carry important implications, as follows: (1) short-term crypto forecasting faces fundamental predictability limits, (2) architectural complexity provides negative ROI in efficient markets, and (3) rigorous statistical validation reveals that apparent improvements often represent noise rather than signal. Full article

As a globally significant oilseed and food crop, peanuts exhibit significant quality changes influenced by storage conditions. This study monitored six key quality indicators—including fatty acid content, carbonyl content, peroxide value, acid value, phenylacetaldehyde and moisture content—in peanut samples stored for 30 weeks under varying temperature and humidity conditions. A Deep Clustering Network (DCN) was employed for quality grading, yielding superior results compared to Deep Empirical Correlation (DEC) and K-Means++ clustering methods, thereby establishing effective quality grading standards. Building upon this, a D-SCSformer time series prediction model was constructed to forecast quality indicators. Through dimensionality-segmented embedding and statistical feature fusion, it achieved strong predictive performance (MSE = 0.2012, MAE = 0.2884, RMSE = 0.4387, and R² = 0.9998), reducing MSE by 57.9%, MAE by 35.4%, and RMSE by 34.1%, while improving R² from 0.9996 to 0.9998 compared to the mainstream Crossformer model. This study provides technical support and a decision-making basis for temperature and humidity regulation and shelf-life management during peanut storage. Full article

The expansion of renewables in modern power systems and the coordinated development of upstream and downstream industrial chains are promoting a shift on the utility side from traditional settlement by energy toward operation driven by data and models. Industrial electricity consumption data exhibit pronounced multi-scale temporal structures and sectoral heterogeneity, which makes unified long-term load and generation forecasting while maintaining accuracy, interpretability, and scalability a challenge. From a modern system identification perspective, this paper proposes a System Identification in Adaptive Bayesian Framework (SIABF) for medium- and long-term industrial load forecasting based on daily freeze electricity time series. By combining daily aggregation of high-frequency data, frequency domain analysis, sparse identification, and long-term extrapolation, we first construct daily freeze series from 15 min measurements, and then we apply discrete Fourier transforms and a spectral complexity index to extract dominant periodic components and build an interpretable sinusoidal basis library. A sparse regression formulation with ${\ell }_1$ regularization is employed to select a compact set of key basis functions, yielding concise representations of sector and enterprise load profiles and naturally supporting multivariate and joint multi-sector modeling. Building on this structure, we implement a state-space-implicit physics-informed Bayesian forecasting model and evaluate it on real data from three representative sectors, namely, steel, photovoltaics, and chemical, using one year of 15 min measurements. Under a one-month-ahead evaluation using one year of 15 min measurements, the proposed framework attains a Mean Absolute Percentage Error (MAPE) of 4.5% for a representative PV-related customer case and achieves low single-digit MAPE for high-inertia sectors, often outperforming classical statistical models, sparse learning baselines, and deep learning architectures. These results should be interpreted as indicative given the limited time span and sample size, and broader multi-year, population-level validation is warranted. Full article

Urban rail transit networks in developing countries are rapidly expanding, entering a networked operational phase where accurate passenger flow forecasting is crucial for optimizing vehicle scheduling, resource allocation, and transportation efficiency. In the short term, accurate real-time forecasting enables the dynamic adjustment of train headways and crew deployment, reducing average passenger waiting times during peak hours and alleviating platform overcrowding; in the long term, reliable trend predictions support strategic planning, including capacity expansion, station retrofitting, and energy management. This paper proposes a novel hybrid forecasting model, EEMD-TiDE, that combines improved Ensemble Empirical Mode Decomposition (EEMD) with a Time Series Dense Encoder (TiDE) to enhance prediction accuracy. The EEMD algorithm effectively overcomes mode mixing issues in traditional EMD by incorporating white noise perturbations, decomposing raw passenger flow data into physically meaningful Intrinsic Mode Functions (IMFs). At the same time, the TiDE model, a linear encoder–decoder architecture, efficiently handles multi-scale features and covariates without the computational overhead of self-attention mechanisms. Experimental results using Xi’an Metro passenger flow data (2017–2019) demonstrate that EEMD-TiDE significantly outperforms baseline models. This study provides a robust solution for urban rail transit passenger flow forecasting, supporting sustainable urban development. Full article

Bike-sharing use has been growing because it improves personal mobility, offers an alternative to walking, and strengthens connections to transit. Demand forecasting is crucial for bike-sharing services because it enables operators to anticipate empty stations and full docks, improve vehicle rebalancing and staffing, and deliver more reliable service at lower operating cost. In this paper, we propose a cluster-based, hour-ahead demand forecasting methodology that (1) groups stations into geographically coherent areas using K-means clustering method, (2) constructs hourly arrival and departure demand time series for each cluster while explicitly preserving zero-demand hours, and (3) incorporates exogenous factors such as temperature and weather-event type. We analyze multi-year trip records from Chicago’s Divvy bike-sharing system (2014–2017) to characterize network expansion and assess spatial stability over time. We then use the period (1 August 2016–31 December 2017), during which the number of active stations is stable, to conduct our predictive modeling. We compare three machine learning-based predictive models—linear regression (LR), time series (TS), and random forest (RF)—and assess their out-of-sample performance using the root mean squared error (RMSE). Results show that TS and RF models consistently outperform LR, achieving up to 80% $R^2$ values and substantially lower RMSE across all 10 clusters, with particular improvements in high-variability central areas. By forecasting net demand (arrivals minus departures) at the cluster level, the approach supports practical identification of likely surplus/deficit areas to guide rebalancing decisions. Full article

Power load forecasting is critical for ensuring grid security and stability and optimizing energy resource allocation. The high integration of renewable energy poses significant challenges to traditional methods in data-scarce scenarios. Recently, Large Language Models (LLMs) have shown considerable potential in processing time-series data, yet their effectiveness in very short-term power load forecasting lacks systematic evaluation. This paper proposes a targeted prompt engineering framework and conducts a systematic empirical study on various LLMs, including GPT-4, Claude-3, Gemini, the Llama series, DeepSeek, and Qwen, comparing them with traditional methods such as ARIMA, BiLSTM, MICN, TimesNet, and VMD-BiLSTM. Furthermore, Ele-LLM, a specialized model based on the Low-Rank Adaptation (LoRA) parameter-efficient fine-tuning strategy, is proposed. Experimental results show that Ele-LLM achieves the best forecasting performance (MAPE = 1.04%), significantly outperforming the best traditional baseline. LLMs also demonstrate notable advantages in few-shot learning, long-sequence dependency modeling, and generalization in complex scenarios. This study provides an evaluation benchmark and practical guidelines for applying LLMs in very short-term power load forecasting, proving their great potential and practical value as an emerging technological pathway. Full article

Background: Ongoing healthcare and medical education reforms in Kazakhstan have been accompanied by persistent workforce shortages and reduced inpatient capacity in pediatric care. Therefore, this study aimed to assess and forecast selected healthcare system indicators using acute mastoiditis (AM) as a sentinel condition while also describing its clinical and epidemiological characteristics. Materials and Methods: This study combined an analysis of national healthcare and demographic statistics in Kazakhstan from 1998 to 2024 with a retrospective review of pediatric AM patients treated at a tertiary referral center. Long-term trends in healthcare resources were assessed, and future needs were projected via average annual percentage change (AAPC) and time series forecasting methods. Clinical, laboratory, and radiological data were extracted from medical records. Statistical analyses were performed via SPSS version 24.0 (IBM Corp., Armonk, NY, USA). Results: From 1998 to 2024, the number of pediatricians and ENT hospital beds declined, whereas the density of ENT physicians remained relatively stable, and the proportion of ENT surgical procedures increased. Projections to 2030 suggest continued constraints in pediatric and ENT workforce capacity and further reductions in inpatient beds despite sustained growth in surgical demand. Among 95 pediatric AM cases, complications, most commonly subperiosteal abscess and zygomatic abscess, were identified in 40% of patients. Conclusions: AM may be considered a contextual indicator of pressures within specialized pediatric ENT services rather than a direct measure of healthcare system performance. These findings highlight the need for further studies to validate these observations and better inform healthcare planning. Full article

High-precision time-series forecasting allows companies to better allocate resources, improve their competitiveness, and increase revenues. In most real-world cases, however, time series are highly volatile and cannot be used for forecasting together with classical statistical methods, which usually yield errors of around 30% or even more. Thus, the goal of this work is to present an approach to obtaining day-ahead forecasts of electricity consumption based on such volatile time series, along with data preprocessing for volatility attenuation. For a thorough understanding, predictions were computed using various methods based on either Artificial Intelligence or purely statistical algorithms. The architectures based on the Transformer were optimized through Brute Force, while the N-BEATS architecture was optimized with Brute Force and OPTUNA because of the highly stochastic nature of the time series. The best method was based on an Encoder-only Transformer, which resulted in an approximate prediction error of 11.63%—far below the error of about 30% usually accepted in current practice. In addition, a procedure was developed to determine the maximum theoretical Pearson Correlation Coefficient between forecast and actual power demand. Full article

Accurate spatiotemporal forecasting underpins high-stakes decision making in smart urban systems, from traffic control and energy scheduling to environment monitoring. Yet two persistent gaps limit current models: (i) spatial modules are often biased toward low-pass smoothing and struggle to reconcile slow global trends with sharp local dynamics; and (ii) the graph structure required for forecasting is frequently latent, while learned graphs can be unstable when built from temporally derived node features alone. We propose SpeQNet, a query-enhanced spectral graph filtering framework that jointly strengthens node representations and graph construction while enabling frequency-selective spatial reasoning. SpeQNet injects global spatial context into temporal embeddings via lightweight learnable spatiotemporal queries, learns a task-oriented adaptive adjacency matrix, and refines node features with an enhanced ChebNetII-based spectral filtering block equipped with channel-wise recalibration and nonlinear refinement. Across twelve real-world benchmarks spanning traffic, electricity, solar power, and weather, SpeQNet achieves state-of-the-art performance and delivers consistent gains on large-scale graphs. Beyond accuracy, SpeQNet is interpretable and robust: the learned spectral operators exhibit a consistent band-stop-like frequency shaping behavior, and performance remains stable across a wide range of Chebyshev polynomial orders. These results suggest that query-enhanced spatiotemporal representation learning and adaptive spectral filtering form a complementary and effective foundation for effective spatiotemporal forecasting. Full article

The dynamics of open quantum systems play a crucial role in quantum information science. However, obtaining numerically exact solutions for the Lindblad master equation is often computationally expensive. Recently, machine learning techniques have gained considerable attention for simulating open quantum system dynamics. In this paper, we propose a deep learning model based on time series prediction (TSP) to forecast the dynamical evolution of open quantum systems. We employ the positive operator-valued measure (POVM) approach to convert the density matrix of the system into a probability distribution and construct a TSP model based on Transformer neural networks. This model effectively captures the historical evolution patterns of the system and accurately predicts its future behavior. Our results show that the model achieves high-fidelity predictions of the system’s evolution trajectory in both short- and long-term scenarios, and exhibits robust generalization under varying initial states and coupling strengths. Moreover, we successfully predicted the steady-state behavior of the system, further proving the practicality and scalability of the method. Full article

This study evaluates the performance of several machine learning models in predicting dissolved oxygen concentration in the surface layer of the Mar Menor coastal lagoon. In recent years, this ecosystem has suffered a continuous process of eutrophication and episodes of hypoxia, mainly due to continuous influx of nutrients from agricultural activities, causing severe water quality deterioration and mortality of local flora and fauna. In this context, monitoring the ecological status of the Mar Menor and its watershed is essential to understand the environmental dynamics that trigger these dystrophic crises. Using field data, this study evaluates the performance of eight predictive modelling approaches, encompassing regularised linear regression methods (Ridge, Lasso, and Elastic Net), instance-based learning (k-nearest neighbours, KNN), kernel-based regression (support vector regression with a radial basis function kernel, SVR-RBF), and tree-based ensemble techniques (Random Forest, Regularised Random Forest, and XGBoost), under multiple experimental settings involving spatial variability and varying time lags applied to physicochemical and meteorological predictors. The results showed that incorporating time lags of approximately two weeks in physicochemical variables markedly improves the models’ ability to generalise to new data. Tree-based regression models achieved the best overall performance, with eXtreme Gradient Boosting providing the highest evaluation metrics. Finally, analysing predictions by sampling point reveals spatial patterns, underscoring the influence of local conditions on prediction quality and the need to consider both spatial structure and temporal inertia when modelling complex coastal lagoon systems. Full article

The time-scale mismatch between rapid influent fluctuations and slow biochemical responses hinders the stability of wastewater treatment plants (WWTPs). Existing models often fail to capture shock signals due to noise interference (“signal pollution”). To address this, we propose the HD-MAED-LSTM model, which employs a “decompose-and-conquer” strategy. Targeting the dynamic characteristics of different components, this study innovatively designs heterogeneous attention mechanisms: utilizing Long-term Dependency Attention to capture the global evolution of the trend component, employing Multi-scale Periodic Attention to reinforce the cyclic patterns of the seasonal component, and using Gated Anomaly Attention to keenly capture sudden shocks in the residual component. In a case study, the effectiveness of the proposed model was validated based on one year of operational data from a large-scale industrial WWTP. HD-MAED-LSTM outperformed baseline models such as Transformer and LSTM in the medium-to-long-term (10-h) prediction of COD, TN, and TP, clearly demonstrating the positive role of differentiated modeling. Notably, in the core task of shock load early warning, the model achieved an F1-Score of 0.83 (superior to Transformer’s 0.77 and LSTM’s 0.67), and a Mean Directional Accuracy (MDA) as high as 0.93. Ablation studies confirm that the specialized attention mechanism is the key performance driver, reducing the Mean Absolute Error (MAE) by 56.7%. This framework provides precise support for shifting WWTPs from passive response to proactive control. Full article

We propose the Hybrid Spatio-Temporal Transformer with Multi-Branch CNN/RNN and Extreme-Event Head (HST–MB–CREH), a hybrid spatio-temporal deep learning architecture for joint short-term photovoltaic (PV) power forecasting and the detection of rare extreme events, to support the reliable operation of renewable-rich power systems. The model combines a spatio-temporal transformer encoder with three convolutional neural network (CNN)/recurrent neural network (RNN) branches (CNN → long short-term memory (LSTM), LSTM → gated recurrent unit (GRU), CNN → GRU) and a dense pathway for tabular meteorological and calendar features. A multitask output head simultaneously performs the regression of PV power and binary classification of extremes defined above the 95th percentile. We evaluate HST–MB–CREH on the publicly available Renewable Power Generation and Weather Conditions dataset with hourly resolutions from 2017 to 2022, using a 5-fold TimeSeriesSplit protocol to avoid temporal leakage and to cover multiple seasons. Compared with tree ensembles (RandomForest, XGBoost), recurrent baselines (Stacked GRU, LSTM), and advanced hybrid/transformer models (Hybrid Multi-Branch CNN–LSTM/GRU with Dense Path and Extreme-Event Head (HMB–CLED) and Spatio-Temporal Multitask Transformer with Extreme-Event Head (STM–EEH)), the proposed architecture achieves the best overall trade-off between accuracy and rare-event sensitivity, with normalized performance of RMSE_z = 0.2159 ± 0.0167, MAE_z = 0.1100 ± 0.0085, mean absolute percentage error (MAPE) = 9.17 ± 0.45%, R² = 0.9534 ± 0.0072, and AUC_ext = 0.9851 ± 0.0051 across folds. Knowledge extraction is supported via attention-based analysis and permutation feature importance, which highlight the dominant role of global horizontal irradiance, diurnal harmonics, and solar geometry features. The results indicate that hybrid spatio-temporal multitask architectures can substantially improve both the forecast accuracy and robustness to extremes, making HST–MB–CREH a promising building block for intelligent decision-support tools in smart grids with a high share of PV generation. Full article

NFT prices are shaped by heterogeneous signals including visual appearance, textual narratives, transaction trajectories, and on-chain interactions, yet existing studies often model these factors in isolation and rarely unify multimodal alignment, temporal non-stationarity, and heterogeneous relational dependencies in a leakage-safe forecasting setting. We propose MM-Temporal-Graph, a cross-modal temporal graph transformer framework for explainable NFT valuation and information-centric risk forecasting. The model encodes image, text, transaction time series, and blockchain behavioral features, constructs a heterogeneous NFT interaction graph (co-transaction, shared creator, wallet relation, and price co-movement), and jointly performs relation-aware graph attention and global temporal–structural transformer reasoning with an adaptive fusion gate. A contrastive multimodal alignment objective improves robustness under market drift, while a risk-aware regularizer and a multi-source risk index enable early warning and interpretable attribution across modalities, time segments, and relational neighborhoods. On MultiNFT-T, MM-Temporal-Graph improves MAE from 0.162 to 0.153 and R² from 0.823 to 0.841 over the strongest multimodal graph baseline, and achieves 87.4% early risk detection accuracy. These results support accurate, robust, and explainable NFT valuation and proactive risk monitoring in Web3 markets. Full article