Classifying time series using feature extraction
Switch branches/tags
Nothing to show
Clone or download
Latest commit 31026cd Oct 9, 2018

README.md

Time Series Classification

Install tsfresh (pip install tsfresh).

Edit config.py for the dataset you want to handle. By default, it's Wafer. Two more are provided in the data\ directory: Ford A and Ford B. You can use any dataset from the UEA & UCR Time Series Classification Repository.

When ready, run

  1. extract_features.py
  2. select_features.py
  3. train_and_evaluate.py

Step one takes some time, so you can skip it - each dataset directory already contains extracted features.

The code uses Python 2 (print statements).