Subset Random Sampling of Finite Time-vertex Graph Signals

Hang Sheng,Qinji Shu,Hui Feng,Bo Hu

2024-10-30

Abstract:Time-varying data with irregular structures can be described by finite time-vertex graph signals (FTVGS), which represent potential temporal and spatial relationships among multiple sources. While sampling and corresponding reconstruction of FTVGS with known spectral support are well investigated, methods for the case of unknown spectral support remain underdeveloped. Existing random sampling schemes may acquire samples from any vertex at any time, which is uncommon in practical applications where sampling typically involves only a subset of vertices and time instants. In sight of this requirement, this paper proposes a subset random sampling scheme for FTVGS. We first randomly select some rows and columns of the FTVGS to form a submatrix, and then randomly sample within the submatrix. Theoretically, we prove sufficient conditions to ensure that the original FTVGS is reconstructed with high probability. Also, we validate the feasibility of reconstructing the original FTVGS by experiments.

Signal Processing

What problem does this paper attempt to address?

The problem that this paper attempts to solve is: **How to perform effective subset random sampling and reconstruction on finite time - vertex graph signals (FTVGS) in the case of unknown joint time - vertex Fourier transform (JFT) spectral support**.

Specifically, existing methods have well - studied the sampling and reconstruction problems of FTVGS when the spectral support is known. However, in practical applications, it is usually impossible to know the JFT spectral support of FTVGS in advance. For example, in sensor networks, specific FTVGS information cannot be predicted before deploying sensors. Therefore, sampling methods based on known spectra are not feasible in practical applications.

To address this challenge, the author proposes a new subset random sampling scheme, which randomly selects samples only from a sub - matrix of FTVGS instead of sampling in the entire matrix. This method not only meets practical requirements but also can significantly reduce the number of sensors that need to be sampled. The author proves through theoretical analysis that under certain conditions, the original FTVGS can be reconstructed with high probability and verifies the effectiveness of this method through experiments.

### Main contributions
1. **Proposed a subset random sampling scheme**: This scheme first randomly selects some rows and columns of FTVGS to form a sub - matrix, and then performs random sampling within this sub - matrix.
2. **Theoretical guarantee**: The author proves that under common assumptions, the proposed subset random sampling scheme can reconstruct the original FTVGS with high probability.
3. **Experimental verification**: It is proved through experiments that the method is effective on actual data sets, especially its superior performance at low sampling rates.

### Key formulas
- Subset random sampling conditions:
\[
|I| \geq 3r\mu_1(X) \ln \left(\frac{2r}{\delta}\right) / \epsilon^2, \quad |J| \geq 3r\mu_2(X) \ln \left(\frac{2r}{\delta}\right) / \epsilon^2
\]
where \( r \) is the rank of matrix \( X \), \(\mu_1(X)\) and \(\mu_2(X)\) are the incoherence parameters of matrix \( X \), \(\delta\) is the failure probability, and \(\epsilon\) is the error range.

- Reconstruction conditions:
\[
|S| \geq 32\beta (\kappa(X))^4 r^2 N / ((1 - \eta)^3 \mu_1(X) \mu_2(X) (|I| + |J|)) \log^2(2n)
\]
where \( S \) is the sampling set, \(\beta > 1\), \(n=\max \{|I|, |J|\}\), \(\kappa(X)\) is the condition number of matrix \( X \), and \(\eta \in [0, 1)\).

Through these formulas and theoretical analysis, the author provides a solid theoretical basis for subset random sampling and shows its feasibility in practical applications.