On the distribution of the entries of a fixed-rank random matrix over a finite field (Q6185645)

Let \(\mathbb{F}_q\) be a finite field of \(q\) elements. For every matrix \(M\) over \(\mathbb{F}_q\), let \(\mathrm{wt}(M)\) be the weight of \(\mathbb{F}_q\), that is, the number of nonzero entries of \( M\). \textit{T. Migler} et al. [Math. Mag. 79, No. 4, 262--271 (2006; Zbl 1134.15002)] proved a formula for the expected value of \(\mathrm{wt}(M)\) when \(M\) is taken at random, with uniform distribution, from the set of \(m \times n\) rank-\(r\) matrices over \(\mathbb{F}_q\). In addition, they suggested that, as \(m, n\) tend to \(+\infty\) and \(r, q\) are fixed, an appropriate scaling of \(\mathrm{wt}(M)\) approaches a normal distribution. The author [Finite Fields Appl. 87, Article ID 102157, 7 p. (2023; Zbl 1512.15043)] proved this last claim for \(q = 2\) and assuming that \(m/n\) converges to a positive real number. Let \(A\) be a nonempty proper subset of \(\mathbb{F}_q\). Moreover, let \(M\) be a random \(m\times n\) rank-\(r\) matrix over \(\mathbb{F}_q\) taken with uniform distribution. In this paper the author proves, in a precise sense, that, as \(m, n\) tend to \(+\infty\) and \(r, q\), \(A\) are fixed, the number of entries of \(M\) that belong to \(A\) approaches a normal distribution taken with uniform distribution. Roughly speaking, this result asserts that, as \(m\) and \(n\) both grow, the number of entries of \(M\) that belong to \(A\) approaches a normal random variable with some prescribed values for its average and variance. Notice that, if the condition on the rank is dropped, that is, if \(M\) is taken at random with uniform distribution from the set of \(m \times n\) matrices over \(\mathbb{F}_q\), then an easy application of the central limit theorem yields that the number of entries of \(M\) that belong to \(A\) approaches a normal random variable with some prescribed values for its average and variance.