Score-based change point detection via tracking the best of infinitely many experts
About
We propose an algorithm for nonparametric online change point detection based on sequential score function estimation and the tracking the best expert approach. The core of the procedure is a version of the fixed share forecaster tailored to the case of infinite number of experts and quadratic loss functions. The algorithm shows promising results in numerical experiments on artificial and real-world data sets. Its performance is supported by rigorous high-probability bounds describing behaviour of the test statistic in the pre-change and post-change regimes.
Anna Markovich, Nikita Puchkin• 2024
Related benchmarks
| Task | Dataset | Result | Rank | |
|---|---|---|---|---|
| Change Point Detection | Synthetic data Example 1 (test) | FA Rate0.00e+0 | 5 | |
| Change Point Detection | Synthetic data Example 2 (test) | FA0.00e+0 | 5 | |
| Change Point Detection | Synthetic data Example 3 (test) | FA Rate0.00e+0 | 5 | |
| Change Point Detection | Synthetic data Example 4 (test) | False Alarm Rate0.00e+0 | 5 | |
| Change Point Detection | CENSREC-1-C Clean Record | FA Rate0.00e+0 | 5 | |
| Change Point Detection | CENSREC-1-C SNR 20 | FA0.00e+0 | 5 | |
| Change Point Detection | CENSREC-1-C SNR 15 | FA0.00e+0 | 5 | |
| Human activity detection | WISDM | FA Rate1 | 5 | |
| Change Point Detection | Room Occupancy (test) | False Alarm Rate1 | 5 |
Showing 9 of 9 rows