Time Series Foundation Models (TSFMs) have emerged as a promising approach in time series analysis. Due to the large-scale parameters of TSFMs and pretraining cost, how to adapt TDFMs in streaming data is always the key factor constraining their application effectiveness. Because streaming data often experiences data distribution and task drifts, which cannot be learnt by offline training. Existing methods typically address streaming data modeling with continuous learning through model fine-tuning or model editing. However, fine-tuning incurs significant computational costs, while editing methods can lead to shifts in the original feature space during streaming updates. To address these limitations, we propose a novel Orthogonal Rotation Transformation-based Continuous Learning method, called ORTCL, for TSFMs. Our key insight is to apply orthogonal matrix rotations to the input and output feature spaces of the TSFMs during model editing. This preserves the metric structure of the original feature space and enables new data to be directly mapped into the existing feature space of the TSFMs. Specifically, we obtain the orthogonal matrix for the input layer via singular value decomposition and derive the corresponding transformation matrix for the output layer through least squares optimization. Extensive experimental results demonstrate that ORTCL outperforms existing methods in both single-domain and cross-domain streaming time series forecasting tasks, effectively mitigating catastrophic forgetting.

Customer value prediction is an essential task for effective customer relationship management, particularly for business-to-business (B2B) service platforms. In B2B scenarios, customer value is jointly shaped by customer preferences and platform service characteristics. Prior research mainly focuses on modeling pairwise relationships between individual customers, failing to capture the multifaceted factors influencing customer value on B2B platforms. To fill this gap, we propose a Multi-view Hypergraph Convolutional Network with Counterfactual Optimization (MHCC) to achieve key-factor-aware customer value prediction. First, to capture hierarchical high-order relationships among business customers, we propose a residual multi-view hypergraph convolutional module. It leverages distinct hypergraphs to model customers' industry categories and service interactions. An attention-based residual fusion layer integrates these relational features with historical behavioral features. Second, we propose a counterfactual pruning optimization module that employs counterfactual perturbation and reasoning to quantify the causal influence of individual connections in the hypergraph. This enables the pruning of redundant connections and the identification of key factors driving customer value. Extensive experiments on a large-scale dataset from a real-world B2B platform demonstrate the superior performance of our model, achieving an 8.59 % improvement in MAE over state-of-the-art methods. We further identify distinctive key factors for customers at different value tiers, thereby offering actionable insights for data-driven customer relationship management.

With the development of information technology, time series plays a vital role in various industries and fields such as finance, transportation, and e-commerce. In recent years, with the in-depth study of time series prediction models based on machine learning, two questions have been further proposed. One is that complex temporal patterns need more comprehensive feature extraction methods to capture the multiscale trends in time series data. Second, deep learning models are treated as black boxes, which makes it usually unconvincing to explain the prediction results. Therefore, this paper proposes an interpretable time series prediction method based on multi-scale feature extraction. In-depth periodic analysis of the input one-dimensional time series data is carried out to reveal its inherent periodicity and trend, and the multi-scale time series features are effectively integrated to generate accurate predictions. By introducing auxiliary functional models into the prediction module, the interpretability of the prediction process is enhanced. Finally, the experimental results on four real-world datasets show that the proposed model exceeds the results of the baseline model in terms of various evaluation metrics and can provide an explanation of the prediction process from both trend and seasonal perspectives.

Spatio-temporal prediction focuses on jointly modeling spatial correlations and temporal evolution and has a wide range of applications. Due to the heterogeneity of spatio-temporal data, accurate prediction relies on effectively integrating topological structures and sequential patterns. Although recurrent graph learning methods excel at capturing dynamic graph patterns, explicitly inferring future snapshots from historical dynamic graphs remains a significant challenge. Recently, prompt-based graph learning has shown the potential to improve future snapshot inference by leveraging node or task-specific prompts. However, these methods fail to fully capture edge information resulting in incomplete and less accurate representations of future snapshot structures. To bridge this gap, we propose ProST, a framework that Prompts future snapshots on dynamic graphs for Spatio-Temporal prediction, which leverages dynamic graph pre-training to generate a premise graph containing historical graph information and then employs prompts on the premise graph to infer explicit future snapshots. Specifically, this framework comprises three steps: Firstly, dynamic graph pre-training is performed using multi-granularity evolution graph convolution to obtain the premise graph with both local and global features of dynamic graphs. Secondly, prompt subgraphs are used to prompt node pairs and edge features within the premise graph. The subgraph prompt aggregation mechanism propagates this information to generate future snapshots. Finally, we freeze the parameters of the pre-trained model and update the subgraph prompt parameters using meta-learning to adapt to downstream spatio-temporal prediction tasks. Extensive experiments on real-world datasets validate that ProST achieves state-of-the-art performance.

With the advent of intelligent technology, the users of spatio-temporal crowdsourcing and their participation in the crowdsourcing tasks continue to increase exponentially. This poses new challenges to the crowdsourcing field. One of the core research areas of spatio-temporal crowdsourcing is task assignment. Most of the existing research on task assignment is focused on offline optimal task assignment, where, the platform has already learned all the information about workers and tasks beforehand. However, these studies cannot obtain good results in real-world situations. At the same time, online task assignment problems often result in local optimal assignment. To solve these problems, more attention needs to be paid to online task assignments and the arrival time of workers. This paper proposes an Online Bilateral Assignment (OBA) problem based on the online assignment model. The competitive ratio of the Greedy algorithm is analyzed according to the OBA problem model. Also, another solution to the OBA problem according to the Greedy algorithm, the Improved-Baseline algorithm, is proposed. Additionally, a Bilateral Online Priority Reassignment algorithm (BOPR) is proposed. The BOPR algorithm realizes real-time task/worker assignment through the bilateral assignment as a solution for online task assignment. In order to guarantee the number of matching tasks, a priority queue is designed in the BOPR algorithm. Considering the waiting time deadlines of tasks and workers and the error rate for priority ranking, it avoids tasks and workers waiting too long and assigns each task to the best possible extent. On this basis, a two-stage assignment strategy is designed for unsuccessful tasks, which could minimize the error rate of the task and significantly improve the efficiency of task assignment. Finally, through experiments on real data sets, the algorithm's performance in terms of global utility value and the number of matches is evaluated.