Title:Top eigenvalue statistics of diluted Wishart matrices

Abstract:Using the replica method, we compute analytically the average largest eigenvalue of diluted covariance matrices of the form $\mathbf{J} = \mathbf{X}^T \mathbf{X}$, where $\mathbf{X}$ is a $N\times M$ sparse data matrix, in the limit of large $N,M$ with fixed ratio. We allow for random non-zero weights, provided they lead to an isolated largest eigenvalue. By formulating the problem as the optimisation of a quadratic Hamiltonian constrained to the $N$-sphere at low temperatures, we derive a set of recursive distributional equations for auxiliary probability density functions, which can be efficiently solved using a population dynamics algorithm. The average largest eigenvalue is identified with a Lagrange parameter that governs the convergence of the algorithm. We find excellent agreement between our analytical results and numerical results obtained from direct diagonalisation.