2026· Proceedings of the 15th International Conference on Data Science, Technology and Applications· 0 citations· 32 references
TL;DR
This work studies multi-horizon taxi demand prediction as the task of learning a mapping from past observations to multiple future demand values under temporal dependence and uncertainty, and suggests that successful multi-horizon forecasting requires global temporal interaction and explicit quantification of uncertainty.
Abstract
: Accurate short-term taxi demand forecasting is essential for efficient urban transportation, as it directly influences fleet allocation, passenger waiting times, and quality of service. This work studies multi-horizon taxi demand prediction as the task of learning a mapping from past observations to multiple future demand values under temporal dependence and uncertainty. A data-driven framework is developed using data from the New York City Taxi and Limousine Commission, where raw trip records are transformed into structured temporal representations through lag-based features, cyclical encoding, and statistical summaries. The framework evaluates sequence modeling architectures for multi-step prediction, such as recurrent models, in particular Long Short-Term Memory (LSTM) networks, and attention-based models, i.e., Transformer architectures. Evaluation across multiple forecasting horizons reveals that attention-based models consistently achieve lower prediction error, indicating their capability to capture the long-term temporal dependencies needed to predict multiple horizons. Conformal prediction and Monte Carlo Dropout are used to produce calibrated prediction intervals and point forecasts that represent demand variability under uncertainty. The findings indicate that successful multi-horizon forecasting requires global temporal interaction and explicit quantification of uncertainty. The proposed framework provides a systematic and scalable approach for real-world deployment in intelligent transportation systems.
Tourist flow modeling and prediction in scenic areas is highly complex, exhibiting significant spatiotemporal dependence and being influenced by various external factors. Traditional models struggle to simultaneously characterize complex spatiotemporal relationships and integrate diverse external information, posing challenges for intelligent resource management and dynamic information scheduling in large-scale networked systems. To address this, this paper proposes a closed-loop integration framework that combines a particle swarm optimization (PSO)-based Temporal Fusion Transformer-Graph Attention Network (TFT-GAT) prediction model with Deep Double-Q Network (D3QN) scheduling optimization. The framework integrates heterogeneous data such as historical traffic, weather, and social media, achieving adaptive time-varying spatial embedding through a graph attention network (GAT) while capturing long- and short-term dependencies using a Temporal Fusion Transformer (TFT) for both point and quantile prediction. The PSO algorithm performs global optimization of the TFT-GAT hyperparameters, and the resulting prediction outputs together with uncertainty estimates are incorporated into the D3QN to realize closed-loop online capacity allocation based on reinforcement learning. Such a data-driven spatiotemporal modeling strategy also provides methodological insights for dynamic information fusion and adaptive resource scheduling in intelligent electromagnetic sensing and communication environments. Experiments conducted at the Forbidden City in Beijing demonstrate high prediction accuracy with a mean absolute error of 1.5–2.6 people/hour, an average quantile coverage exceeding 82%, and a response time of 140.6 ms. The incorporation of exogenous factors, particularly holidays, improves prediction performance by 18.4%, validating the proposed framework’s robustness in spatiotemporal coupling, uncertainty representation, and real-time scheduling.
X. N. Xie, W. L. Zhang· Advanced Electromagnetics· 0 citations
An intelligent framework is developed as a hybrid one, where four complementary learners are used to operate in parallel: a spatio-temporal graph neural network (ST-GNN) to represent dependencies between travel zones, a Transformer to represent long-horizon temporal patterns, an LSTM to represent sequential mobility dynamics, and an XGBoost to represent structured demographic and cost characteristics.
Santosh Kumar Sharma, Satish Chander, Piyush Gupta· International journal of com...· 0 citations
An integrated prediction-and-visualisation pipeline that transforms complex data distributions into actionable visual analytics, such as interpretable station-to-station demand heatmaps via interactive GIS Folium layers is implemented, providing an operationally robust framework to support smart-city transportation management and build more sustainable urban transit systems.
Berna Çalışkan· Journal of Data Analytics an...· 0 citations
TAccurate short-term electricity consumption forecasting is essential for operational planning, reserve allocation, and energy management in modern power systems. This study investigates the performance of recurrent and hybrid attention-based deep learning architectures for short-term electricity consumption forecasting, a sase study in Togo. The proposed framework integrates electricity consumption, meteorological, demographic, and temporal information in order to capture both intrinsic temporal dependencies and exogenous influences affecting electricity demand. Four forecasting architectures were evaluated using a weekly temporal window of 168 hours: LSTM-only, LSTM-decoder, LSTM-attention, and a hybrid LSTM--Multi-Head Attention--LSTM model. The experiments included multi-seed evaluation, ablation study, robustness analysis, and statistical comparison using the Wilcoxon signed-rank test. The results show that all models achieved extremely high forecasting accuracy, with coefficients of determination exceeding 0.9998 and MAPE values below 0.004%. The hybrid architectures slightly improved average forecasting performance, while the standalone LSTM model remained highly competitive. The robustness analysis revealed strong sensitivity to noisy inputs but moderate degradation under missing-data conditions. Statistical analysis indicated that the performance differences between architectures were not statistically significant at the 5% level. Overall, the study demonstrates that recurrent deep learning architectures provide reliable and operationally relevant solutions for electricity consumption forecasting in highly structured energy demand series.
Séna Apeke, Yao Bokovi, K. Gbafa et al.· Science Journal of Energy En...· 0 citations
Short-horizon demand forecasting for automotive spare parts remains a challenging task due to sparse observations, frequent zero-demand periods, and limited historical data at the customer-product level. While neural forecasting models have shown promising predictive capabilities, their practical adoption in operational environments is often constrained by limited transparency and trust. In such contexts, understanding model behavior is as critical as achieving high forecast accuracy. This study presents an explainable neural forecasting framework designed to support short-horizon demand prediction under data scarcity. The approach combines zero-inflated modelling, hierarchical representations, and attention-based temporal mechanisms to capture recent demand dynamics while enabling controlled information sharing across related series. Rather than focusing exclusively on numerical performance, the framework emphasizes interpretability through post-hoc feature attribution and temporal relevance analysis. The proposed model is evaluated on a real-world automotive spare parts dataset using benchmark forecasting methods as references. Model behavior is analyzed through explainability tools and illustrative diagnostics, including temporal attention patterns and residual analyses, to assess transparency and stability. The results demonstrate that the framework effectively captures short-term demand signals and provides interpretable insights into the factors influencing predictions. These findings highlight the potential of explainable neural forecasting models as decision-support tools for spare parts management in data-scarce industrial settings.
Gizem Erdinç· 2026 6th International Confe...· 0 citations
Accurate residential user cluster load forecasting is essential for advancing
demand-side management in smart distribution grids and assisting power grid companies in
peak–valley regulation. To overcome the limitations of traditional load forecasting methods, this
study combines multi-temporal segmentation with the Mamba (TS-Mamba) architecture to develop
a high-precision forecasting framework.
The historical load sequence is first segmented into several temporal segments. A convolution-
based perceptron subsequently processes these segments to capture local features, while
temporal dependencies are modeled through a selective state-space formulation. A gated perceptron
structure is further integrated to improve feature interaction and model robustness. In addition,
residual connections inspired by ResNet are introduced to stabilize network training.
Experimental results across different load datasets show that the proposed method consistently
outperforms conventional neural network and Transformer-based forecasting models. Moreover,
the parameter-efficient state-space structure, combined with temporal segmentation and perceptron-
based feature learning, enables the model to maintain stable prediction performance even
in small-sample scenarios.
The proposed model relies on historical load patterns for learning temporal dependencies.
Therefore, its performance may degrade when the load series contains abrupt structural
changes, extreme events, or highly irregular patterns that are not sufficiently represented in the
training data. In addition, the segmentation strategy introduces hyperparameters that may require
careful tuning for different datasets.
The proposed model effectively improves the accuracy of short-term load forecasting.
It is worth noting that in the small-sample scenario, TS-Mamba can maintain reliable predictive
performance while reducing the number of model parameters.
Ye Zhang, Xiaoyu Zhang, Ailiang Kang et al.· Recent Advances in Electrica...· 0 citations