The current code is running based on a patient's data (ID=540) in the Ohio dataset.

You should have a subdirectory called "processedcsv" with "ohio540.csv" in it.

To run the code:

python main.py --data ohio540

This project contains the Pytorch implementation of the following paper:

Title: A Novel Hyperdimensional Computing Framework for Online Time Series Forecasting on the Edge

We present a novel framework for efficient time series forecasting in edge computing. It departs from traditional, resource-intensive deep learning methods, embracing hyperdimensional computing (HDC) for a more efficient approach. The framework includes two models: the Autoregressive Hyperdimensional Computing (AR-HDC) and the Sequence-to-Sequence HDC (Seq2Seq-HDC). These models are designed to reduce inference times and improve accuracy in both short-term and long-term forecasting, making them ideal for resource-limited edge computing scenarios

• Python 3.10.12
• numpy==1.26.3
• pandas==2.1.4
• scikit-learn==1.3.2
• torch==2.1.2
• tqdm==4.66.1

Please install the required packages listed in the requirements.txt file using the following command :

pip install -r requirements.txt

We follow the same data formatting as the Informer repo (https://github.com/zhouhaoyi/Informer2020), which also hosts the raw data. Please put all raw data (csv) files in the ./data folder.

To replicate our results on the ETT, ECL, Exchange, Illness, and WTH datasets, run

chmod +x scripts/*.sh
bash .scripts/run.sh

Method: Our implementation supports the following training strategies:

• AR-HDC: Autoregressive Hyperdimensional Computing Framework
• Seq2Seq-HDC: Sequence-to-Sequence Hyperdimensional Computing Framework

You can specify one of the above method via the --method argument.

Dataset: Our implementation currently supports the following datasets: Electricity Transformer - ETT (including ETTh1, ETTh2, ETTm1, and ETTm2), ECL, Exchange, Illness and WTH. You can specify the dataset via the --data argument.

Other arguments: Other useful arguments for experiments are:

• --hvs_len: Dimension of the hyperspace: e.g. D = 1000 ,
• --seq_len: look-back windows' length, set to 2 * τ by default,
• --pred_len: forecast windows' &tau length

If you find this repository useful in your research, please consider citing the following papers:

@misc{title={A Novel Hyperdimensional Computing Framework for Online Time Series Forecasting on the Edge}, 
      year={2024},
      eprint={2402.01999},
      archivePrefix={arXiv},
      primaryClass={cs.LG}
}