{"id":13526140,"url":"https://github.com/ztxtech/Time-Evidence-Fusion-Network","last_synced_at":"2025-04-01T06:31:17.693Z","repository":{"id":239509997,"uuid":"790225590","full_name":"ztxtech/Time-Evidence-Fusion-Network","owner":"ztxtech","description":"Official implementation of \"Time Evidence Fusion Network: Multi-source View in Long-Term Time Series Forecasting\" (https://arxiv.org/abs/2405.06419)","archived":false,"fork":false,"pushed_at":"2024-10-21T10:52:32.000Z","size":3685,"stargazers_count":65,"open_issues_count":0,"forks_count":4,"subscribers_count":2,"default_branch":"master","last_synced_at":"2024-11-02T10:34:11.509Z","etag":null,"topics":["cuda","deep-learning","machine-learning","macos","neural-network","neural-networks","pytorch","time-series","time-series-analysis","time-series-forecasting","time-series-prediction","uestc"],"latest_commit_sha":null,"homepage":"https://ztxtech.github.io/Time-Evidence-Fusion-Network/","language":"Python","has_issues":true,"has_wiki":null,"has_pages":null,"mirror_url":null,"source_name":null,"license":"mit","status":null,"scm":"git","pull_requests_enabled":true,"icon_url":"https://github.com/ztxtech.png","metadata":{"files":{"readme":"README.md","changelog":null,"contributing":null,"funding":null,"license":"LICENSE","code_of_conduct":null,"threat_model":null,"audit":null,"citation":null,"codeowners":null,"security":null,"support":null,"governance":null,"roadmap":null,"authors":null,"dei":null,"publiccode":null,"codemeta":null}},"created_at":"2024-04-22T13:48:06.000Z","updated_at":"2024-10-31T04:32:29.000Z","dependencies_parsed_at":"2024-05-20T12:47:22.771Z","dependency_job_id":"836d51b4-152f-4c84-af77-78205173948f","html_url":"https://github.com/ztxtech/Time-Evidence-Fusion-Network","commit_stats":null,"previous_names":["ztxtech/time-evidence-fusion-network"],"tags_count":0,"template":false,"template_full_name":null,"repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/ztxtech%2FTime-Evidence-Fusion-Network","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/ztxtech%2FTime-Evidence-Fusion-Network/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/ztxtech%2FTime-Evidence-Fusion-Network/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/ztxtech%2FTime-Evidence-Fusion-Network/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/ztxtech","download_url":"https://codeload.github.com/ztxtech/Time-Evidence-Fusion-Network/tar.gz/refs/heads/master","host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":246596839,"owners_count":20802905,"icon_url":"https://github.com/github.png","version":null,"created_at":"2022-05-30T11:31:42.601Z","updated_at":"2022-07-04T15:15:14.044Z","host_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub","repositories_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories","repository_names_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repository_names","owners_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners"}},"keywords":["cuda","deep-learning","machine-learning","macos","neural-network","neural-networks","pytorch","time-series","time-series-analysis","time-series-forecasting","time-series-prediction","uestc"],"created_at":"2024-08-01T06:01:25.679Z","updated_at":"2025-04-01T06:31:15.206Z","avatar_url":"https://github.com/ztxtech.png","language":"Python","readme":"\u003cdiv align=\"center\"\u003e\n  \u003ch2\u003e\u003cb\u003e Time Evidence Fusion Network (TEFN): \n    \u003cbr/\u003e Multi-source View in Long-Term Time Series Forecasting \u003c/b\u003e\u003c/h2\u003e\n\u003c/div\u003e\n\n**Repo Status:**\n\n![PRs Welcome](https://img.shields.io/badge/PRs-Welcome-green) \n[![Visits Badge](https://badges.pufler.dev/visits/ztxtech/Time-Evidence-Fusion-Network)](https://github.com/ztxtech/Time-Evidence-Fusion-Network)\n[![GitHub last commit](https://img.shields.io/github/last-commit/ztxtech/Time-Evidence-Fusion-Network)](https://github.com/ztxtech/Time-Evidence-Fusion-Network/activity?ref=master\u0026activity_type=direct_push)\n[![GitHub commit activity](https://img.shields.io/github/commit-activity/t/ztxtech/Time-Evidence-Fusion-Network)](https://github.com/ztxtech/Time-Evidence-Fusion-Network/graphs/commit-activity)\n[![GitHub code size in bytes](https://img.shields.io/github/languages/code-size/ztxtech/Time-Evidence-Fusion-Network)](https://github.com/ztxtech/Time-Evidence-Fusion-Network)\n[![GitHub Repo stars](https://img.shields.io/github/stars/ztxtech/Time-Evidence-Fusion-Network)](https://github.com/ztxtech/Time-Evidence-Fusion-Network)\n[![GitHub forks](https://img.shields.io/github/forks/ztxtech/Time-Evidence-Fusion-Network)](https://github.com/ztxtech/Time-Evidence-Fusion-Network)\n[![GitHub watchers](https://img.shields.io/github/watchers/ztxtech/Time-Evidence-Fusion-Network)](https://github.com/ztxtech/Time-Evidence-Fusion-Network)\n\n**Implementation:**\n\n[![arxiv](https://img.shields.io/badge/cs.LG-2405.06419-b31b1b?style=flat\u0026logo=arxiv\u0026logoColor=red)](https://arxiv.org/abs/2405.06419)\n[![Python](https://img.shields.io/badge/python-3670A0?logo=python\u0026logoColor=ffdd54)](https://www.python.org/)\n[![PyTorch](https://img.shields.io/badge/PyTorch-%23EE4C2C.svg?logo=PyTorch\u0026logoColor=white)](https://pytorch.org/)\n[![nVIDIA](https://img.shields.io/badge/nVIDIA-cuda-%2376B900.svg?logo=nVIDIA\u0026logoColor=white)](https://pytorch.org/docs/2.1/cuda.html)\n[![Apple](https://img.shields.io/badge/Mac-MPS-%23000000.svg?logo=apple\u0026logoColor=white)](https://pytorch.org/docs/2.1/mps.html)\n\n## Updates\n\n🚩 **News** (2024.05.14) Compatible with MPS backend, TEFN can be trained by [![Apple](https://img.shields.io/badge/MacBook_Air_2020-M1_8G-%23000000.svg?logo=apple\u0026logoColor=white)](https://support.apple.com/zh-cn/111883).\n\n\n## Overview\n\nThis is the official code implementation project for paper **\"Time Evidence Fusion Network: Multi-source View in\nLong-Term Time Series Forecasting\"**. The code implementation refers\nto [![GitHub](https://img.shields.io/badge/thuml-Time_Series_Library-%23121011?logo=github\u0026logoColor=white)](https://github.com/thuml/Time-Series-Library).\nThanks very much\nfor [![GitHub](https://img.shields.io/badge/thuml-Time_Series_Library-%23121011?logo=github\u0026logoColor=white)](https://github.com/thuml/Time-Series-Library)'s contribution to this project.\n\n![TEFN](./fig/TEFN.png)\nThe **Time Evidence Fusion Network (TEFN)** is a groundbreaking deep learning model designed for long-term time series\nforecasting. It integrates the principles of information fusion and evidence theory to achieve superior performance in\nreal-world applications where timely predictions are crucial. TEFN introduces the Basic Probability Assignment (BPA)\nModule, leveraging fuzzy theory, and the Time Evidence Fusion Network to enhance prediction accuracy, stability, and\ninterpretability.\n\n## Key Features\n\n- **Information Fusion Perspective**: TEFN addresses time series forecasting from a unique angle, focusing on the fusion\n  of multi-source information to boost prediction accuracy.\n  ![Information Fusion Perspective](./fig/ms.png)\n- **BPA Module**: At its core, TEFN incorporates a BPA Module that maps diverse information sources to probability\n  distributions related to the target outcome. This module exploits the interpretability of evidence theory, using fuzzy\n  membership functions to represent uncertainty in predictions.\n  ![BPA Diagram](./fig/inver_conv.png)\n  ![BPA](./fig/bpa.png)\n- **Interpretability**: Due to its roots in fuzzy logic, TEFN provides clear insights into the decision-making process,\n  enhancing model explainability.\n  ![Channel dimension interpretability](./fig/CBV.png)\n  ![Time dimension interpretability](./fig/TBV.png)\n- **State-of-the-Art Performance**: TEFN demonstrates competitive results, with prediction errors comparable to leading\n  models like PatchTST, while maintaining high efficiency and requiring fewer parameters than complex models such as\n  Dlinear.\n  ![SOTA](./fig/sota.png)\n- **Robustness and Stability**: The model showcases resilience to hyperparameter tuning, exhibiting minimal fluctuations\n  even under random selections, ensuring consistent performance across various settings.\n  ![Visualization of Robustness](./fig/vr.png)\n  ![Variance](./fig/var.png)\n- **Efficiency**: With optimized training times and a compact model footprint, TEFN is particularly suitable for\n  resource-constrained environments.\n  ![Efficiency](./fig/size.png)\n\n## Getting Started\n\n### Requirements\n\n- ![Python](https://img.shields.io/badge/python-\u003e3.6-3670A0?logo=python\u0026logoColor=ffdd54) Python \u003e= 3.6\n- ![PyTorch](https://img.shields.io/badge/PyTorch-\u003e1.7.0-%23EE4C2C.svg?logo=PyTorch\u0026logoColor=white) PyTorch \u003e= 1.7.0\n- ![Python](https://img.shields.io/badge/PyPI-3670A0?logo=PyPI\u0026logoColor=ffdd54) Other dependencies listed\n  in `requirements.txt`\n\n### Installation\n\nClone the repository:\n\n```bash\ngit clone https://github.com/ztxtech/Time-Evidence-Fusion-Network.git\ncd Time-Evidence-Fusion-Network\npip install -r requirements.txt\n```\n\n### Usage\n\n#### Download Dataset\n\nYou can obtain datasets\nfrom [![Google Drive](https://img.shields.io/badge/Google%20Drive-4285F4?logo=googledrive\u0026logoColor=white)](https://drive.google.com/drive/folders/13Cg1KYOlzM5C7K8gK8NfC-F3EYxkM3D2?usp=sharing)\nor [![Baidu Drive](https://img.shields.io/badge/Baidu-Pan-2932E1?logo=Baidu\u0026logoColor=white)](https://pan.baidu.com/s/1r3KhGd0Q9PJIUZdfEYoymg?pwd=i9iy),\nThen place the downloaded data in the folder`./dataset`.\n\n#### Load Config\n\n1. Modify the specific configuration file in `./run_config.py`.\n\n```python\nconfig_path = '{your chosen config file path}'\n```\n\n2. Run `./run_config.py` directly.\n\n```bash\npython run_config.py\n```\n\n#### Switching Running Devices\n\n1. Find required configuration file `*.json` in `./configs`.\n2. Modify `*.json` file.\n\n``` \n{\n  # ...\n  # Nvidia CUDA Device {0}\n  # 'gpu': 0\n  # Apple MPS Device\n  # 'gpu': 'mps'\n  # ...\n}\n```\n\n#### Other Operations\n\nOther related operations refer\nto [![GitHub](https://img.shields.io/badge/thuml-Time_Series_Library-%23121011?logo=github\u0026logoColor=white)](https://github.com/thuml/Time-Series-Library).\n\n#### Citation\n\nIf you find TEFN useful in your research, please cite our work as per the citation.\n\n```bibtex\n@misc{TEFN,\n      title={Time Evidence Fusion Network: Multi-source View in Long-Term Time Series Forecasting}, \n      author={Tianxiang Zhan and Yuanpeng He and Zhen Li and Yong Deng},\n      year={2024},\n      journal={arXiv}\n}\n\n```\n\n## Acknowledgement\n\nWe appreciate the following GitHub repos a lot for their valuable code and efforts.\n\n- [Time Series Library ![GitHub](https://img.shields.io/badge/thuml-Time_Series_Library-%23121011?logo=github\u0026logoColor=white)](https://github.com/thuml/Time-Series-Library)'\n- [TSFpaper ![GitHub](https://img.shields.io/badge/ddz16-TSFpaper-%23121011?logo=github\u0026logoColor=white)](https://github.com/ddz16/TSFpaper)\n- [Time-Series-Forecasting-and-Deep-Learning ![GitHub](https://img.shields.io/badge/DaoSword-Time--Series--Forecasting--and--Deep--Learning-%23121011?logo=github\u0026logoColor=white)](https://github.com/DaoSword/Time-Series-Forecasting-and-Deep-Learning)\n- [awesome-opensource ![GitHub](https://img.shields.io/badge/gitroomhq-awesome--opensource-%23121011?logo=github\u0026logoColor=white)](https://github.com/gitroomhq/awesome-opensource)\n\n### From [Time Series Library](https://github.com/thuml/Time-Series-Library)\n\n\nThis library is constructed based on the following repos:\n\n- Forecasting: https://github.com/thuml/Autoformer.\n\nAll the experiment datasets are public, and we obtain them from the following links:\n\n- Long-term Forecasting and Imputation: https://github.com/thuml/Autoformer.\n\n- Short-term Forecasting: https://github.com/ServiceNow/N-BEATS.\n\n\n\n## Contact\n\nIf you have any questions or suggestions, feel free to contact:\n\n- (**Primary**) Tianxiang Zhan [(ztxtech@std.uestc.edu.cn)](mailto:ztxtech@std.uestc.edu.cn)\n  [![Outlook](https://img.shields.io/badge/Tianxiang_Zhan-0078D4?logo=microsoft-outlook\u0026logoColor=white)](mailto:ztxtech@std.uestc.edu.cn)\n  [![Google Scholar](https://img.shields.io/badge/Tianxiang_Zhan-4285F4?logo=googlescholar\u0026logoColor=white)](https://scholar.google.com.hk/citations?user=bRYz250AAAAJ)\n  [![ResearchGate](https://img.shields.io/badge/Tianxiang_Zhan-00CCBB?logo=ResearchGate\u0026logoColor=white)](https://www.researchgate.net/profile/Tianxiang-Zhan)\n- Yuanpeng He [(heyuanpeng@stu.pku.edu.cn)](mailto:heyuanpeng@stu.pku.edu.cn)\n  [![Outlook](https://img.shields.io/badge/Yuanpeng_He-0078D4?logo=microsoft-outlook\u0026logoColor=white)](mailto:heyuanpeng@stu.pku.edu.cn)\n  [![Google Scholar](https://img.shields.io/badge/Yuanpeng_He-4285F4?logo=googlescholar\u0026logoColor=white)](https://scholar.google.com/citations?user=HaefBCQAAAAJ)\n  [![ResearchGate](https://img.shields.io/badge/Yuanpeng_He-00CCBB?logo=ResearchGate\u0026logoColor=white)](https://www.researchgate.net/profile/Yuanpeng-He)\n\nOr describe it in Issues.\n\n## Star History\n\n[![Star History Chart](https://api.star-history.com/svg?repos=ztxtech/Time-Evidence-Fusion-Network\u0026type=Date)](https://star-history.com/#ztxtech/Time-Evidence-Fusion-Network\u0026Date)\n","funding_links":[],"categories":["Python","Papers"],"sub_categories":["2024"],"project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fztxtech%2FTime-Evidence-Fusion-Network","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Fztxtech%2FTime-Evidence-Fusion-Network","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fztxtech%2FTime-Evidence-Fusion-Network/lists"}