https://github.com/huawei-noah/BHT-ARIMA
Code for paper: Block Hankel Tensor ARIMA for Multiple Short Time Series Forecasting (AAAI-20)
https://github.com/huawei-noah/BHT-ARIMA
arima-forecasting tensor-decomposition tensor-factorization time-series
Last synced: about 1 month ago
JSON representation
Code for paper: Block Hankel Tensor ARIMA for Multiple Short Time Series Forecasting (AAAI-20)
- Host: GitHub
- URL: https://github.com/huawei-noah/BHT-ARIMA
- Owner: huawei-noah
- License: mit
- Created: 2020-02-25T02:43:54.000Z (about 5 years ago)
- Default Branch: master
- Last Pushed: 2021-06-08T01:29:04.000Z (almost 4 years ago)
- Last Synced: 2025-03-23T23:26:45.566Z (about 2 months ago)
- Topics: arima-forecasting, tensor-decomposition, tensor-factorization, time-series
- Language: Python
- Size: 61.5 KB
- Stars: 105
- Watchers: 3
- Forks: 40
- Open Issues: 4
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
- awesome-time-series - Block Hankel Tensor ARIMA for Multiple Short Time Series Forecasting
README
# BHT-ARIMA
A tensor decomposition-based time series forecasting algorithm, which tactically incorporates the unique advantages of Hankelization, low-rank Tucker decomposition and ARIMA into a unified framework.
More details (including parameter settings) refer to [the original paper](https://arxiv.org/abs/2002.12135).
### Paper
- **"[Block Hankel Tensor ARIMA for Multiple Short Time Series Forecasting](https://arxiv.org/abs/2002.12135)", AAAI-20**
### Datasets
Traffic dataset. The traffic data is originally collected from California department of transportation 1 and describes the road occupy rate of Los Angeles County highway network.We here use the same subset used in (Yu, Yin, and Zhu 2017) which selects **228 sensors** randomly. And We take **the first 40 time points** of them as data of our demo
### Getting Started
#### Prerequisites
- python 3.5+
- python libraries
- tensorly
- scipy
- numpy
- pandas
#### Run
```python
python main.py
```
### License
© Contributors, 2019. Licensed under an [MIT](LICENSE) license.