Predictive Model Team Project

• Libraries Used: https://nixtlaverse.nixtla.io/

File Description

• TSAD_JY[Final_X].ipynb -- Univariate with Exogenous TSAD (Multivariate)
  Experimental Models: LSTM, GRU, NBEATSx, NHITS, TSMixerx

• TSAD_JY[Final_M].ipynb -- Multivariate TSAD
  Experimental Models: TSMixer, NHITS, PatchTST, TimesNet

Key Parameters

• horizon: Prediction length
• input_size: Training window size
• threshold: Anomaly detection threshold

RUN CODE

ENVIRONMENT SET
Ubuntu 20.04 LTS CUDA 12.1.0 cudnn8

conda create -n TS PYTHON==3.9
pip install pip
pip install neuralforecast

Choose what you want :)

bash TSAD_JY.sh

OR

If you want to give your own parameters, see below:

python TSAD_JY_X.py --horizon 10 --input_size 20 threshold 4.0
python TSAD_JY_M.py --horizon 20 --input_size 30 threshold 2.0