Unverified50% confidenceFactExact time
In their 2004 paper, Ledoit and Wolf derived an analytically optimal solution for the shrinkage intensity δ that minimizes the Frobenius norm between the estimated matrix and the true covariance matrix.
1
Sources
50%
Confidence
Long-term
Relevance
9/11/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedLedoit and Wolf proposed the Analytical Nonlinear Shrinkage method in 2020, which performs optimal shrinkage on eigenvalues one by one65% similarUnverifiedEckart-Young-Mirsky定理证明，在所有秩为k的矩阵中，由SVD前k个奇异值构成的截断近似矩阵是在Frobenius范数意义下对原矩阵的最佳近似64% similarUnverifiedThe mathematical expression for Ledoit-Wolf shrinkage is Σ_shrunk = δ·F + (1-δ)·S, where S is the sample covariance matrix, F is the shrinkage target, and δ∈[0,1] is the shrinkage intensity.61% similarUnverified指针分析经典算法中，Andersen 分析精度更高但计算复杂度达 O(n³)，Steensgaard 分析近乎线性但精度较低60% similarUnverified哈达玛猜想由法国数学家雅克·哈达玛于 1893 年提出，认为当矩阵阶数是 4 的倍数时哈达玛矩阵一定存在，至今未被完全证明60% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/891761API
curl https://kongchang.com/api/v1/knowledge/claims/891761MCP
get_claim(id=891761)