https://github.com/adityalab/camul
https://github.com/adityalab/camul
Last synced: 28 days ago
JSON representation
- Host: GitHub
- URL: https://github.com/adityalab/camul
- Owner: AdityaLab
- License: mit
- Created: 2021-09-17T15:40:07.000Z (over 3 years ago)
- Default Branch: master
- Last Pushed: 2022-02-18T16:50:26.000Z (about 3 years ago)
- Last Synced: 2025-04-02T23:51:32.033Z (29 days ago)
- Language: Python
- Size: 552 MB
- Stars: 14
- Watchers: 2
- Forks: 10
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
# CaMUL: Calibrated and Accurate Multi-view Time-Series Forecasting
## Setup
We require you to have `anaconda` or `miniconda` installed. Run the script `./scripts/setup.sh` to setup the virtual environment with all the required packages.
## Data download and preprocessing
### Twitter dataset
The probability distributions for each week over all states are available in `./data/tweet_dataset` folder as npy files for each week and state.
### Power dataset
Run the `./scripts/download_power.sh` to download dataset.
### Covid dataset
Covid dataset is available in `./data/covid_data` folder. Run `./scripts/covid_preprocess.sh` to preprocess the features of dataset.
### Google Symptoms
Symptoms dataset is available in `./data/symptom_data`. Run `./scripts/preprocess_symp.sh` for preprocessing.
## Experiments
We have `./train_tweets.py`, `./train_covid.py`, `./train_power.py`, `./train_symp.py` to run the model for each of the benchmarks. You may tune the arguments related week ahead, prediction week/season by passing the commandline arguments. Use the `--help` flag for a list of all arguments.
## Acknowledgement
In case you use our code or datasets, please cite us as:
```
@article{kamarthi2021camul,
title={CAMul: Calibrated and Accurate Multi-view Time-Series Forecasting},
author={Kamarthi, Harshavardhan and Kong, Lingkai and Rodr{\'\i}guez, Alexander and Zhang, Chao and Prakash, B Aditya},
journal={Proceedings of the Web Conference 2022},
year={2022}
}
```