tire-cpd 0.1.3

This is an unsupervised change point detection toolbox, including time-invariant representation (TIRE) model with diamond loss, multi-channel time-invariant representation (MC-TIRE) model, and multi-view time-invariant representation (multiview TIRE) model.

TIRE-cpd toolbox

Toolbox for time-invariant representation autoencoder approach (TIRE) for change point detection (CPD) task. Including three models: TIRE model with diamond loss [1], multi-channel TIRE model [2], and multi-view TIRE model [3].

The authors of these papers are:

• Zhenxiang Cao (STADIUS, Dept. Electrical Engineering, KU Leuven)
• Nick Seeuws (STADIUS, Dept. Electrical Engineering, KU Leuven)
• Maarten De Vos (STADIUS, Dept. Electrical Engineering, KU Leuven and Dept. Development and Regeneration, KU Leuven)
• Alexander Bertrand (STADIUS, Dept. Electrical Engineering, KU Leuven)

All authors are affiliated to LEUVEN.AI - KU Leuven Institute for AI.

Use examples and function explanations can be found in GitHub repository: tire-cpd_toolbox_example.

References

• [1] Cao, Z., Seeuws, N., De Vos, M. and Bertrand, A., 2023. A novel loss for change point detection models with time-invariant representations. IEEE Signal Processing Letters, 30, pp.1737-1741.
• [2] Cao, Z., Seeuws, N., De Vos, M. and Bertrand, A., 2023. Change Point Detection in Multi-Channel Time Series Via a Time-Invariant Representation. IEEE Transactions on Knowledge and Data Engineering.
• [3] Cao, Z., Seeuws, N., De Vos, M. and Bertrand, A., 2024. A Multi-view Extension for Change Point Detection via Time-invariant Representations. Proceedings of EUSIPCO 2024.