Explainable time-series machine learning using the tsfresh module

The central idea of this project is the observation that publications using systematic time series feature extraction from the Open Source machine learning library tsfresh, always need to visualize and explain the extracted features. Due to the fact that tsfresh uses 168 algorithms to compute up to 800 time series features, this process can be highly automated using an object oriented approach. This project will contribute towards the development of a Python module for explainable time-series machine learning, which will assists researchers and engineers in the interpretation of their time-series based machine learning models.

Faculty Supervisor:

Arthur Chan

Student:

Partner:

University of Auckland

Discipline:

Engineering

Sector:

Artificial Intelligence

University:

University of Toronto

Program: