Complete monotonicity of some entropies (Q1705827)
From MaRDI portal
scientific article
Language | Label | Description | Also known as |
---|---|---|---|
English | Complete monotonicity of some entropies |
scientific article |
Statements
Complete monotonicity of some entropies (English)
0 references
16 March 2018
0 references
It is well-known that the Shannon entropies of some parameterized probability distributions are concave functions with respect to the parameter. The main aim of this paper is to consider a family of such distributions (including the binomial, Poisson, and negative binomial distributions) and investigate their Shannon, Rényi, and Tsallis entropies with respect to complete monotonicity. Let \(c\in \mathbb{R}\) and \( I_{c}= [0, -\frac{1}{c}] \text{ if} c<0\) and \(I_{c}= [0, +\infty[, \text{ if} c\geq 0\). Let \(n > 0\) be a real number such that \(n > c\) if \(c \geq 0\), or \(n = -c l\) with some \(l \in \mathbb{N}\) if \(c < 0\). For \(k\in \mathbb{N}_{0}\) and \(x\in I_{c}\) define \[ p_{n, k}^{[c]}(x)= (-1)^{k}\left(k^{-\frac{n}{c}}\right)(cx)^{k}(1+cx)^{-\frac{n}{c}-k} \; \left(c\neq 0\right) \] \[ p_{n, k}^{[0]}(x)\lim_{c\to 0}p_{n, k}^{[c]}(x)=\dfrac{(nx)^{k}}{k!}e^{-nx}. \] The Shannon entropy associated to the parameterized probability distribution \(\left(p_{n, k}^{[c]}(x)\right)_{k\in \mathbb{N}_{0}}\) is \[ H_{n, c}(x)= -\sum_{k=0}^{\infty}p_{n, k}^{[c]}(x)\log \left(p_{n, k}^{[c]}(x)\right). \] Furthermore, the Rényi entropy of order \(2\) and the Tsallis entropy of order \(2\) are given by \[ R_{n, c}(x)= -\log \left(S_{n, c}(x)\right) \] and \[ T_{n, c}(x)=1-S_{n, c}(x), \] respectively, where \[ S_{n, c}(x)= \sum_{k=0}^{\infty}\left(p_{n, k}^{[c]}(x)\right)^{2}. \] At first, in Theorem 2.1, the author determined the signs of all derivative of \(H_{n, c}\) in case \(c<0\). After that he shows that in case \(c=0\) the function \(H'_{n, 0}\) is completely monotonic, that is, \[ (-1)^{k}H^{(k+1)}_{n, 0}(x)\geq 0 \] for all \(x>0\) and for any \(k\geq 0\). This is Theorem 2.2 of the paper. Finally, in the last result of the second section, he derives that in case \(c>0\), \(H'_{n, 0}\) is completely monotonic, too (see Theorem 2.3). The third section of the paper concerns the above defined mapping \(R_{n, c}\) and \(T_{n, c}\). He (among others) the following result is proved. Let \(c\geq 0\). Then \(R_{n, c}\) is increasing and concave; \(T'_{n, c}\) is completely monotonic on \([0, +\infty[\); \(T_{n, c}\) is concave for all \(c\in \mathbb{R}\).
0 references
entropies
0 references
concavity
0 references
complete monotonicity
0 references
inequalities
0 references