Time Evidence Fusion Network: Multi-source View in Long-Term Time Series Forecasting
- URL: http://arxiv.org/abs/2405.06419v1
- Date: Fri, 10 May 2024 12:10:22 GMT
- Title: Time Evidence Fusion Network: Multi-source View in Long-Term Time Series Forecasting
- Authors: Tianxiang Zhan, Yuanpeng He, Zhen Li, Yong Deng,
- Abstract summary: Time series forecasting often demands timeliness, making research on model backbones a perennially hot topic.
We propose a novel backbone from the perspective of information fusion.
In real data experiments, the TEFN partially achieved state-of-the-art, with low errors comparable to PatchTST.
- Score: 10.733698311045181
- License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
- Abstract: In real-world scenarios, time series forecasting often demands timeliness, making research on model backbones a perennially hot topic. To meet these performance demands, we propose a novel backbone from the perspective of information fusion. Introducing the Basic Probability Assignment (BPA) Module and the Time Evidence Fusion Network (TEFN), based on evidence theory, allows us to achieve superior performance. On the other hand, the perspective of multi-source information fusion effectively improves the accuracy of forecasting. Due to the fact that BPA is generated by fuzzy theory, TEFN also has considerable interpretability. In real data experiments, the TEFN partially achieved state-of-the-art, with low errors comparable to PatchTST, and operating efficiency surpass performance models such as Dlinear. Meanwhile, TEFN has high robustness and small error fluctuations in the random hyperparameter selection. TEFN is not a model that achieves the ultimate in single aspect, but a model that balances performance, accuracy, stability, and interpretability.
Related papers
- Fuxi-DA: A Generalized Deep Learning Data Assimilation Framework for Assimilating Satellite Observations [15.934673617658609]
Deep learning models have shown promise in matching, even surpassing, the forecast accuracy of leading NWP models worldwide.
This study introduces FuxiDA, a generalized DL-based DA framework for assimilating satellite observations.
By assimilating data from Advanced Geosynchronous Radiation Imager (AGRI) aboard Fengyun-4B, FuXi-DA consistently mitigates analysis errors and significantly improves forecast performance.
arXiv Detail & Related papers (2024-04-12T15:02:14Z) - Cumulative Distribution Function based General Temporal Point Processes [49.758080415846884]
CuFun model represents a novel approach to TPPs that revolves around the Cumulative Distribution Function (CDF)
Our approach addresses several critical issues inherent in traditional TPP modeling.
Our contributions encompass the introduction of a pioneering CDF-based TPP model, the development of a methodology for incorporating past event information into future event prediction.
arXiv Detail & Related papers (2024-02-01T07:21:30Z) - Parsimony or Capability? Decomposition Delivers Both in Long-term Time Series Forecasting [46.63798583414426]
Long-term time series forecasting (LTSF) represents a critical frontier in time series analysis.
Our study demonstrates, through both analytical and empirical evidence, that decomposition is key to containing excessive model inflation.
Remarkably, by tailoring decomposition to the intrinsic dynamics of time series data, our proposed model outperforms existing benchmarks.
arXiv Detail & Related papers (2024-01-22T13:15:40Z) - TEA: Test-time Energy Adaptation [67.4574269851666]
Test-time adaptation (TTA) aims to improve model generalizability when test data diverges from training distribution.
We propose a novel energy-based perspective, enhancing the model's perception of target data distributions.
arXiv Detail & Related papers (2023-11-24T10:49:49Z) - The Missing U for Efficient Diffusion Models [3.712196074875643]
Diffusion Probabilistic Models yield record-breaking performance in tasks such as image synthesis, video generation, and molecule design.
Despite their capabilities, their efficiency, especially in the reverse process, remains a challenge due to slow convergence rates and high computational costs.
We introduce an approach that leverages continuous dynamical systems to design a novel denoising network for diffusion models.
arXiv Detail & Related papers (2023-10-31T00:12:14Z) - Exploring Progress in Multivariate Time Series Forecasting:
Comprehensive Benchmarking and Heterogeneity Analysis [72.18987459587682]
We introduce BasicTS, a benchmark designed for fair comparisons in MTS forecasting.
We highlight the heterogeneity among MTS datasets and classify them based on temporal and spatial characteristics.
arXiv Detail & Related papers (2023-10-09T19:52:22Z) - Towards Flexible Time-to-event Modeling: Optimizing Neural Networks via
Rank Regression [17.684526928033065]
We introduce the Deep AFT Rank-regression model for Time-to-event prediction (DART)
This model uses an objective function based on Gehan's rank statistic, which is efficient and reliable for representation learning.
The proposed method is a semiparametric approach to AFT modeling that does not impose any distributional assumptions on the survival time distribution.
arXiv Detail & Related papers (2023-07-16T13:58:28Z) - A comparative assessment of deep learning models for day-ahead load
forecasting: Investigating key accuracy drivers [2.572906392867547]
Short-term load forecasting (STLF) is vital for the effective and economic operation of power grids and energy markets.
Several deep learning models have been proposed in the literature for STLF, reporting promising results.
arXiv Detail & Related papers (2023-02-23T17:11:04Z) - Towards Long-Term Time-Series Forecasting: Feature, Pattern, and
Distribution [57.71199089609161]
Long-term time-series forecasting (LTTF) has become a pressing demand in many applications, such as wind power supply planning.
Transformer models have been adopted to deliver high prediction capacity because of the high computational self-attention mechanism.
We propose an efficient Transformerbased model, named Conformer, which differentiates itself from existing methods for LTTF in three aspects.
arXiv Detail & Related papers (2023-01-05T13:59:29Z) - Closed-form Continuous-Depth Models [99.40335716948101]
Continuous-depth neural models rely on advanced numerical differential equation solvers.
We present a new family of models, termed Closed-form Continuous-depth (CfC) networks, that are simple to describe and at least one order of magnitude faster.
arXiv Detail & Related papers (2021-06-25T22:08:51Z) - Transformer Hawkes Process [79.16290557505211]
We propose a Transformer Hawkes Process (THP) model, which leverages the self-attention mechanism to capture long-term dependencies.
THP outperforms existing models in terms of both likelihood and event prediction accuracy by a notable margin.
We provide a concrete example, where THP achieves improved prediction performance for learning multiple point processes when incorporating their relational information.
arXiv Detail & Related papers (2020-02-21T13:48:13Z)
This list is automatically generated from the titles and abstracts of the papers in this site.
This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.