/time-series-classification

Classifying time series using feature extraction

Primary LanguagePythonApache License 2.0Apache-2.0

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).