quick-start.md

In this turorial, we give two examples with UEA & UCR Time Series Classification and Regression Repository.

Classification

Here we take the Heartbeat classification task from UEA & UCR Time Series Classification Repository as an example.

1. Download the dataset

cd PhysioPro
mkdir data
wget http://www.timeseriesclassification.com/aeon-toolkit/Archives/Multivariate2018_ts.zip -P data

unzip data/Multivariate2018_ts.zip -d data/
rm data/Multivariate2018_ts.zip

2. Run Heartbeat classification task with TSRNN model

# create the output directory
mkdir -p outputs/Multivariate_ts/Heartbeat
# run the train task
python -m physiopro.entry.train docs/configs/rnn_classification.yml
# tensorboard
tensorboard --logdir outputs/

The results will be saved to outputs/Multivariate2018_ts/Heartbeat directory.

Regression

Here we take the BeijingPM25Quality dataset from UEA & UCR Time Series Extrinsic Regression Dataset as an example.

1. Download the dataset

mkdir -p data/Monash_UEA_UCR_Regression_Archive/BeijingPM25Quality
wget https://zenodo.org/record/3902671/files/BeijingPM25Quality_TEST.ts?download=1 -O data/Monash_UEA_UCR_Regression_Archive/BeijingPM25Quality/BeijingPM25Quality_TEST.ts
wget https://zenodo.org/record/3902671/files/BeijingPM25Quality_TRAIN.ts?download=1 -O data/Monash_UEA_UCR_Regression_Archive/BeijingPM25Quality/BeijingPM25Quality_TRAIN.ts

2. Run BeijingPM25Quality regression task with TSRNN model

# create the output directory
mkdir -p outputs/Monash_UEA_UCR_Regression_Archive/BeijingPM25Quality
# run the train task with RNN 
python -m physiopro.entry.train docs/configs/rnn_regression.yml
# or with TCN
# python -m physiopro.entry.train docs/configs/tcn_regression.yml
# # or with transformer
# python -m physiopro.entry.train docs/configs/transformer_regression.yml

The results will be saved to outputs/Monash_UEA_UCR_Regression_Archive/BeijingPM25Quality directory.