Abstract:Consider a high-dimensional Wishart matrix $\bd{W}=\bd{X}^T\bd{X}$ where the entries of $\bd{X}$ are i.i.d. random variables with mean zero, variance one, and a finite fourth moment $\eta$. Motivated by problems in signal processing and high-dimensional statistics, we study the maximum of the largest eigenvalues of any two-by-two principal minors of $\bd{W}$. Under certain restrictions on the sample size and the population dimension of $\bd{W}$, we obtain the limiting distribution of the maximum, which follows the Gumbel distribution when $\eta$ is between 0 and 3, and a new distribution when $\eta$ exceeds 3. To derive this result, we first address a simpler problem on a new object named a deformed Gaussian orthogonal ensemble (GOE). The Wishart case is then resolved using results from the deformed GOE and a high-dimensional central limit theorem. Our proof strategy combines the Stein-Poisson approximation method, conditioning, U-statistics, and the Hájek projection. This method may also be applicable to other extreme-value problems. Some open questions are posed.
What problem does this paper attempt to address?
### The problems the paper attempts to solve
What this paper attempts to solve is the problem of the limiting distribution of the largest eigenvalue of the principal minor in high - dimensional Wishart matrices. Specifically, the researchers focus on the maximum of the largest eigenvalues of any two second - order principal minors and explore its limiting distribution under specific sample size and population dimension constraints. The main contribution of the paper is that when the fourth - order moment \(\eta\) is in different intervals, the limiting distribution of the maximum of these largest eigenvalues is found. Specifically:
1. **Background and motivation**:
- Applications in high - dimensional statistics and signal processing have promoted the study of extreme eigenvalues of principal minors of Wishart matrices.
- Researchers are particularly interested in the maximum of the largest eigenvalues of the second - order principal minors of the Wishart matrix \(W = X^T X\), where the elements of \(X\) are independent and identically distributed random variables with a mean of 0, a variance of 1, and a finite fourth - order moment.
2. **Research object**:
- Define \(T_{n,p,2}=\max_{1\leq i < j\leq p}\lambda_1(W_{\{i,j\}})\), where \(W_{\{i,j\}}\) is the second - order principal minor of \(W\).
- The researchers first consider a simpler model - the Deformed Gaussian Orthogonal Ensemble (GOE), and then generalize their results to Wishart matrices through the high - dimensional central limit theorem.
3. **Main results**:
- When the fourth - order moment \(\eta\) is between 0 and 3, the limiting distribution of \(T_{n,p,2}\) is the Gumbel distribution.
- When \(\eta\) exceeds 3, the limiting distribution of \(T_{n,p,2}\) is a new distribution, whose form is given in the paper.
- These results not only verify some previous conjectures but also discover new probability distributions.
4. **Methods**:
- The researchers use techniques such as the Stein - Poisson approximation method, conditioning, U - statistics, and Hájek projection to derive the limiting distribution.
- Through these methods, the researchers successfully transform the problem of Wishart matrices into the problem of the deformed GOE, thus simplifying the analysis.
### Summary
This paper reveals the limiting distributions of the maximum of the largest eigenvalues of the second - order principal minors in high - dimensional Wishart matrices under different fourth - order moment conditions by studying it. This research not only extends the existing theory but also provides new tools and methods for high - dimensional statistics and signal processing.