-
Likelihood distortion and Bayesian local robustness
Authors:
Antonio Di Noia,
Fabrizio Ruggeri,
Antonietta Mira
Abstract:
Robust Bayesian analysis has been mainly devoted to detecting and measuring robustness w.r.t. the prior distribution. Many contributions in the literature aim to define suitable classes of priors which allow the computation of variations of quantities of interest while the prior changes within those classes. The literature has devoted much less attention to the robustness of Bayesian methods w.r.t…
▽ More
Robust Bayesian analysis has been mainly devoted to detecting and measuring robustness w.r.t. the prior distribution. Many contributions in the literature aim to define suitable classes of priors which allow the computation of variations of quantities of interest while the prior changes within those classes. The literature has devoted much less attention to the robustness of Bayesian methods w.r.t. the likelihood function due to mathematical and computational complexity, and because it is often arguably considered a more objective choice compared to the prior. In this contribution, we propose a new approach to Bayesian local robustness, mainly focusing on robustness w.r.t. the likelihood function. Successively, we extend it to account for robustness w.r.t. the prior, as well as the prior and the likelihood jointly. This approach is based on the notion of distortion function introduced in the literature on risk theory. The novel robustness measure is a local sensitivity measure that turns out to be very tractable and easy to compute for several classes of distortion functions. Asymptotic properties are derived, and numerical experiments illustrate the theory and its applicability for modelling purposes.
△ Less
Submitted 2 September, 2025; v1 submitted 23 May, 2024;
originally announced May 2024.
-
Beyond the noise: intrinsic dimension estimation with optimal neighbourhood identification
Authors:
Antonio Di Noia,
Iuri Macocco,
Aldo Glielmo,
Alessandro Laio,
Antonietta Mira
Abstract:
The Intrinsic Dimension (ID) is a key concept in unsupervised learning and feature selection, as it is a lower bound to the number of variables which are necessary to describe a system. However, in almost any real-world dataset the ID depends on the scale at which the data are analysed. Quite typically at a small scale, the ID is very large, as the data are affected by measurement errors. At large…
▽ More
The Intrinsic Dimension (ID) is a key concept in unsupervised learning and feature selection, as it is a lower bound to the number of variables which are necessary to describe a system. However, in almost any real-world dataset the ID depends on the scale at which the data are analysed. Quite typically at a small scale, the ID is very large, as the data are affected by measurement errors. At large scale, the ID can also be erroneously large, due to the curvature and the topology of the manifold containing the data. In this work, we introduce an automatic protocol to select the sweet spot, namely the correct range of scales in which the ID is meaningful and useful. This protocol is based on imposing that for distances smaller than the correct scale the density of the data is constant. In the presented framework, to estimate the density it is necessary to know the ID, therefore, this condition is imposed self-consistently. We derive theoretical guarantees and illustrate the usefulness and robustness of this procedure by benchmarks on artificial and real-world datasets.
△ Less
Submitted 7 September, 2024; v1 submitted 23 May, 2024;
originally announced May 2024.
-
On the mathematical axiomatization of approximate Bayesian computation. A robust set for estimating mechanistic network models through optimal transport
Authors:
Marco Tarsia,
Antonietta Mira,
Daniele Cassani
Abstract:
We research relations between optimal transport theory (OTT) and approximate Bayesian computation (ABC) possibly connected to relevant metrics defined on probability measures. Those of ABC are computational methods based on Bayesian statistics and applicable to a given generative model to estimate its a posteriori distribution in case the likelihood function is intractable. The idea is therefore t…
▽ More
We research relations between optimal transport theory (OTT) and approximate Bayesian computation (ABC) possibly connected to relevant metrics defined on probability measures. Those of ABC are computational methods based on Bayesian statistics and applicable to a given generative model to estimate its a posteriori distribution in case the likelihood function is intractable. The idea is therefore to simulate sets of synthetic data from the model with respect to assigned parameters and, rather than comparing prospects of these data with the corresponding observed values as typically ABC requires, to employ just a distance between a chosen distribution associated to the synthetic data and another of the observed values. Our focus lies in theoretical and methodological aspects, although there would exist a remarkable part of algorithmic implementation, and more precisely issues regarding mathematical foundation and asymptotic properties are carefully analysed, inspired by an in-depth study of what is then our main bibliographic reference, that is Bernton et al. (2019), carrying out what follows: a rigorous formulation of the set-up for the ABC rejection algorithm, also to regain a transparent and general result of convergence as the ABC threshold goes to zero whereas the number n of samples from the prior stays fixed; general technical proposals about distances leaning on OTT; weak assumptions which lead to lower bounds for small values of threshold and as n goes to infinity, ultimately showing a reasonable possibility of lack of concentration which is contrary to what is proposed in Bernton et al. (2019) itself.
△ Less
Submitted 5 May, 2021;
originally announced May 2021.
-
Total Roman 2-domination in graphs
Authors:
Suitberto Cabrera Garcia,
Abel Cabrera Martinez,
Frank A. Hernandez Mira,
Ismael G. Yero
Abstract:
Given a graph $G=(V,E)$, a function $f:V\rightarrow \{0,1,2\}$ is a total Roman $\{2\}$-dominating function if: (1) every vertex $v\in V$ for which $f(v)=0$ satisfies that $\sum_{u\in N(v)}f(u)\geq 2$, where $N(v)$ represents the open neighborhood of $v$, and (2) every vertex $x\in V$ for which $f(x)\geq 1$ is adjacent to at least one vertex $y\in V$ such that $f(y)\geq 1$. The weight of the funct…
▽ More
Given a graph $G=(V,E)$, a function $f:V\rightarrow \{0,1,2\}$ is a total Roman $\{2\}$-dominating function if: (1) every vertex $v\in V$ for which $f(v)=0$ satisfies that $\sum_{u\in N(v)}f(u)\geq 2$, where $N(v)$ represents the open neighborhood of $v$, and (2) every vertex $x\in V$ for which $f(x)\geq 1$ is adjacent to at least one vertex $y\in V$ such that $f(y)\geq 1$. The weight of the function $f$ is defined as $ω(f)=\sum_{v\in V}f(v)$. The total Roman $\{2\}$-domination number, denoted by $γ_{t\{R2\}}(G)$, is the minimum weight among all total Roman $\{2\}$-dominating functions on $G$. In this article we introduce the concepts above and begin the study of its combinatorial and computational properties. For instance, we give several closed relationships between this parameter and other domination related parameters in graphs. In addition, we prove that the complexity of computing the value $γ_{t\{R2\}}(G)$ is NP-hard, even when restricted to bipartite or chordal graphs.
△ Less
Submitted 7 January, 2021;
originally announced January 2021.
-
A note on total co-independent domination in trees
Authors:
Abel Cabrera Martínez,
Frank A. Hernández Mira,
José M. Sigarreta Almira,
Ismael G. Yero
Abstract:
A set $D$ of vertices of a graph $G$ is a total dominating set if every vertex of $G$ is adjacent to at least one vertex of $D$. The total domination number of $G$ is the minimum cardinality of any total dominating set of $G$ and is denoted by $γ_t(G)$. The total dominating set $D$ is called a total co-independent dominating set if $V(G)\setminus D$ is an independent set and has at least one verte…
▽ More
A set $D$ of vertices of a graph $G$ is a total dominating set if every vertex of $G$ is adjacent to at least one vertex of $D$. The total domination number of $G$ is the minimum cardinality of any total dominating set of $G$ and is denoted by $γ_t(G)$. The total dominating set $D$ is called a total co-independent dominating set if $V(G)\setminus D$ is an independent set and has at least one vertex. The minimum cardinality of any total co-independent dominating set is denoted by $γ_{t,coi}(G)$. In this paper, we show that, for any tree $T$ of order $n$ and diameter at least three, $n-β(T)\leq γ_{t,coi}(T)\leq n-|L(T)|$ where $β(T)$ is the maximum cardinality of any independent set and $L(T)$ is the set of leaves of $T$. We also characterize the families of trees attaining the extremal bounds above and show that the differences between the value of $γ_{t,coi}(T)$ and these bounds can be arbitrarily large for some classes of trees.
△ Less
Submitted 2 May, 2020;
originally announced May 2020.
-
On computational and combinatorial properties of the total co-independent domination number of graphs
Authors:
Abel Cabrera Martinez,
Frank A. Hernandez Mira,
Jose M. Sigarreta Almira,
Ismael G. Yero
Abstract:
A subset $D$ of vertices of a graph $G$ is a total dominating set if every vertex of $G$ is adjacent to at least one vertex of $D$. The total dominating set $D$ is called a total co-independent dominating set if the subgraph induced by $V-D$ is edgeless and has at least one vertex. The minimum cardinality of any total co-independent dominating set is the total co-independent domination number of…
▽ More
A subset $D$ of vertices of a graph $G$ is a total dominating set if every vertex of $G$ is adjacent to at least one vertex of $D$. The total dominating set $D$ is called a total co-independent dominating set if the subgraph induced by $V-D$ is edgeless and has at least one vertex. The minimum cardinality of any total co-independent dominating set is the total co-independent domination number of $G$ and is denoted by $γ_{t,coi}(G)$. In this work we study some complexity and combinatorial properties of $γ_{t,coi}(G)$. Specifically, we prove that deciding whether $γ_{t,coi}(G)\le k$ for a given integer $k$ is an NP-complete problem and give several bounds on $γ_{t,coi}(G)$. Also, since any total co-independent dominating set is also a total dominating set, we characterize all the trees having equal total co-independent domination number and total domination number.
△ Less
Submitted 29 August, 2017; v1 submitted 2 May, 2017;
originally announced May 2017.
