A study of distance measures and learning methods for semi-supervised learning on time series data
Other than scipy/numpy, the following distance measures and learning methods have their own prerequesites:
- DTW distance:
dtaidistancefrom [https://github.com/wannesm/dtaidistance] - SoftDTW distance:
sdtwfrom [https://github.com/mblondel/soft-dtw] - Matrix Profile distance:
stumpyfrom [https://github.com/TDAmeritrade/stumpy] - Graph Convolutional Network:
torchandtorch-geometricfrom [https://github.com/rusty1s/pytorch_geometric]
Each of the four methods has its own Python script run-method.py which is controlled by command-line options. Run, e.g., python run-gcn.py --help for a list of arguments.