Skip to content
/ SST Public

[IEEE TSP 2024] "OPIT: A Simple but Effective Sparse Subspace Tracking". In IEEE Transactions on Signal Process. 2024

License

Notifications You must be signed in to change notification settings

thanhtbt/SST

Repository files navigation

OPIT: A Simple but Effective Sparse Subspace Tracking Method

In this work, we propose a new provable effective method called OPIT (which stands for Online Power Iteration via Thresholding) for tracking the sparse principal subspace of data streams over time. Particularly, OPIT introduces a new adaptive variant of power iteration with space and computational complexity linear to the data dimension. In addition, a new column-based thresholding operator is developed to regularize the subspace sparsity. Utilizing both advantages of power iteration and thresholding operation, OPIT is capable of tracking the underlying subspace in both classical and high dimensional regimes.

Demo

Please run

  • demo_effect_forgetting_factor.m: To illustrate the effect of the forgetting factor on the performance of OPIT
  • demo_noise_effect.m: To illustrate the effect of noise on the performance of OPIT
  • demo_nonstationary.m: To illustrate the performance of OPIT in nonstationary environments
  • demo_low_dimension_comparison.m: To illustrate the performance of subspace tracking algorithms in the classical setting
  • demo_high_dimension_comparison.m: To illustrate the performance of subspace tracking algorithms in high dimension

State-of-the-art algorithms for comparison

Some Experimental Results

  • Effect of the sparsity level

    SST_Sparse

  • OPIT vs SOTA Algorithms in High Dimension

SST_Compare

  • Performance of SST algorithms with different data dimensions and sample sizes

SST_Compare_v2

Reference

This code is free and open source for research purposes. If you use this code, please acknowledge the following paper.

[1] L.T. Thanh, K. Abed-Meraim, N.L. Trung, & A. Hafiance. "Sparse Subspace Tracking in High Dimensions". Proc. 47th IEEE ICASSP, 2022.

[2] L.T. Thanh, K. Abed-Meraim, N. L. Trung, & A. Hafiane. "OPIT: A Simple and Effective Method for Sparse Subspace Tracking in High-dimension and Low-sample-size Context". IEEE Trans. Signal Process., 2024.

Releases

No releases published

Packages

No packages published

Languages