Time Series Classification Research

Research repository for developing competitive TSC algorithms. Currently exploring combinations of Quant and Hydra algorithms.

Structure

• quant/, hydra/, aaltd2024/ - Algorithm implementations (submodules)
• tsckit/ - Python package for unified algorithm/dataset interfaces
• *.ipynb - Experiment notebooks
• papers/ - Reference papers

Goal

Build a new time series classification algorithm that competes with state-of-the-art methods. Current focus is on ensemble approaches combining different feature extraction strategies.