arXiv subjects

Lina Oliveira

Publications and source records attributed to Lina Oliveira.

12 recordsLinked to original sources

arXiv2026

Minimum Covariance Determinant Estimator and Outlier Detection for Interval-valued Data

Interval-valued data are one of the most common symbolic data types, which enables the preservation of the underlying variability of the data. The interval mean and covariance matrix can be estimated using the barycenter approach based on the Mallows distance. However, as for conventional data, classical estimates can be significantly affected by anomalous data points, frequently present in real-life datasets. To address this problem, we develop a robust alternative which estimates location and scale by extending the Minimum Covariance Determinant estimator to interval-valued data. The algorithm yields a robust Interval-Mahalanobis distance, which can be used to detect anomalous observations based on adaptive cutoff values. Through extensive simulation studies across various contamination levels, we demonstrate that the interval-valued robust estimator consistently outperforms classical methods in covariance matrix estimation and achieves superior outlier detection accuracy. Finally, the applicability and effectiveness of the proposed method are illustrated through real-world datasets.

stat.ME↗

arXiv2026

Explainable Outlier Detection for Interval-valued Data

Explainability is increasingly recognized as a key aspect of outlier detection. However, for complex data structures such as interval-valued data, it remains largely unexplored. Building on an outlier detection framework based on the Interval Minimum Covariance Determinant estimator, we propose a novel approach to explain the outlyingness of interval-valued observations using the concept of the Shapley value. We derive a closed-form expression for the Shapley value of the squared robust Interval-Mahalanobis distance, enabling efficient computation of variable contributions. This formulation allows for a fine-grained interpretation of outliers, providing a detailed decomposition into contributions from centers, ranges, and cross-terms of the interval-valued observations. Moreover, the Shapley value is closely connected to the concept of cellwise outliers, as it can help identify variable-specific outliers that may not be evident at multivariate level. We further extend the framework through the Shapley interaction index to capture pairwise variable interactions driving atypical behavior. The practical utility of the proposed approach is illustrated through two real-world datasets.

stat.ME↗

arXiv2025

Interval Fisher's Discriminant Analysis and Visualisation

In Data Science, entities are typically represented by single valued measurements. Symbolic Data Analysis extends this framework to more complex structures, such as intervals and histograms, that express internal variability. We propose an extension of multiclass Fisher's Discriminant Analysis to interval-valued data, using Moore's interval arithmetic and the Mallows' distance. Fisher's objective function is generalised to consider simultaneously the contributions of the centres and the ranges of intervals and is numerically maximised. The resulting discriminant directions are then used to classify interval-valued observations.To support visual assessment, we adapt the class map, originally introduced for conventional data, to classifiers that assign labels through minimum distance rules. We also extend the silhouette plot to this setting and use stacked mosaic plots to complement the visual display of class assignments. Together, these graphical tools provide insight into classifier performance and the strength of class membership. Applications to real datasets illustrate the proposed methodology and demonstrate its value in interpreting classification results for interval-valued data.

stat.ML↗

arXiv2025

Location and association measures for interval-valued data based on Mallows' distance

The growing demand to analyse large and complex datasets has spurred the development of Symbolic Data Analysis as a promising approach to address contemporary data challenges. Amongst these, interval-valued data introduces new theoretical and methodological questions that remain open. In this paper, we generalise measures of location and association for interval-valued random variables using Mallows' distance. Departing from restrictive assumptions such as uniform distributions over microdata, our proposal extends the barycentre approach to any absolutely continuous distribution with finite second moment. A key contribution is the derivation of explicit formulas for Mallows' distance in p-dimensional interval spaces. These formulas decompose into components for centres, ranges, and a novel cross-term that captures their interaction. This decomposition leads to a new theoretical symbolic covariance matrix that explicitly accounts for the dependence between centres and ranges - a relation often obscured in current definitions of symbolic covariance. Theoretical developments are supported by empirical studies on diverse real-world datasets, each reflecting different degrees of information about the underlying microdata. These applications highlight both the flexibility of the proposed methodology and the interpretability of its results.

math.ST↗

arXiv2025

Isomorphisms of Tits--Kantor--Koecher Lie algebras of JB*-triples

We characterise the isomorphisms of Tits--Kantor--Koecher Lie algebras of JB*-triples as a class of surjective linear isometries and show how these algebras form a category equivalent to that of JB*-triples. We introduce the concepts of tripotent, and orthogonality and order amongst tripotents for Tits--Kantor--Koecher Lie algebras. This leads to showing that a graded or negatively graded order isomorphism between certain subsets of tripotents of two Tits--Kantor--Koecher Lie algebras of atomic JB*-triples, which commutes with involutions, preserves orthogonality and is continuous at a non-zero tripotent of a specific type, can be extended as a real-linear isomorphism between the algebras.

math.FA↗

arXiv2024

On a class of left ideals of nest algebras

We introduce a class of left ideals (and subalgebras) of nest algebras determined by totally ordered families of partial isometries on a complex Hilbert space $H$. Let $\mathcal{E}$ be a family of partial isometries that is totally ordered in the Halmos--McLaughlin ordering, and let $\mathcal{A}_{\mathcal{E}}$ be the subset of operators in $B(H)$ which, for all $E\in \mathcal{E}$, map the initial space of $E$ to the final space of $E$. We show that $\mathcal{A}_{\mathcal{E}}$ is a subalgebra of $B(H)$ if and only if $\mathcal{A}_{\mathcal{E}}$ is a left ideal of a certain nest algebra, and if so, $\mathcal{E}$ consists of power partial isometries, except possibly for its supremum $\vee \mathcal{E}$, in which case the range $\operatorname{ran}(\vee \mathcal{E})$ is $H$. It is also shown that any left ideal $\mathcal{A}_{\mathcal{E}}$ is decomposable and that the subset of finite rank operators in its closed unit ball is strongly dense in the ball. Necessary and sufficient conditions to solve $Tx=y$ and $T^*x=y$ in $\mathcal{A}_{\mathcal{E}}$ are given.

math.OA↗

arXiv2020

Bimodules of Banach space nest algebras

We extend to Banach space nest algebras the theory of essential supports and support function pairs of their bimodules, thereby obtaining Banach space counterparts of long established results for Hilbert space nest algebras. Namely, given a Banach space nest algebra $\mathcal A$, we charaterise the maximal and the minimal $\mathcal A$-bimodules having a given essential support function or support function pair. These characterisations are complete except for the minimal $\mathcal A$-bimodule corresponding to a support function pair, in which case we make some headway. We also show that the weakly closed bimodules of a Banach space nest algebra are exactly those that are reflexive operator spaces. To this end, we crucially prove that reflexive bimodules determine uniquely a certain class of admissible support functions.

math.FA↗

arXiv2019

Kernel maps and operator decomposition

We introduce the notions of kernel map and kernel set of a bounded linear operator on a Hilbert space relative to a subspace lattice. The characterization of the kernel maps and kernel sets of finite rank operators leads to showing that every norm closed Lie module of a continuous nest algebra is decomposable. The continuity of the nest cannot be lifted, in general.

math.OA↗

arXiv2015

Weakly closed Lie modules of nest algebras

Let $\mathcal{T}(\mathcal{N})$ be a nest algebra of operators on Hilbert space and let $\mathcal{L}$ be a weakly closed Lie $\mathcal{T}(\mathcal{N})$-module. We construct explicitly the largest possible weakly closed $\mathcal{T}(\mathcal{N})$-bimodule $\mathcal{J}(\mathcal{L})$ and a weakly closed $\mathcal{T}(\mathcal{N})$-bimodule $\mathcal{K}(\mathcal{L})$ such that \[ \mathcal{J}(\mathcal{L})\subseteq \mathcal{L} \subseteq \mathcal{K}(\mathcal{L}) +\mathcal{D}_{\mathcal{K}(\mathcal{L})}, \] $[\mathcal{K}(\mathcal{L}), \mathcal{T}(\mathcal{N})]\subseteq \mathcal{L}$ and $\mathcal{D}_{\mathcal{K}(\mathcal{L})}$ is a von Neumann subalgebra of the diagonal $\mathcal{T}(\mathcal{N})\cap \mathcal{T}(\mathcal{N})^*$.

math.OA↗

arXiv2015

A characterization of reflexive spaces of operators

We show that for a linear space of operators ${\mathcal M}\subseteq {\mathcal B}(H_1,H_2)$ the following assertions are equivalent. (i) ${\mathcal M} $ is reflexive in the sense of Loginov--Shulman. (ii) There exists an order-preserving map $Ψ=(ψ_1,ψ_2)$ on a bilattice $Bil({\mathcal M})$ of subspaces determined by ${\mathcal M}$, with $P\leq ψ_1(P,Q)$ and $Q\leq ψ_2(P,Q)$, for any pair $(P,Q)\in Bil({\mathcal M})$, and such that an operator $T\in {\mathcal B}(H_1,H_2)$ lies in ${\mathcal M}$ if and only if $ψ_2(P,Q)Tψ_1(P,Q)=0$ for all $(P,Q)\in Bil( {\mathcal M})$. This extends to reflexive spaces the Erdos--Power type characterization of weakly closed bimodules over a nest algebra.

math.OA↗

arXiv2014

Decomposability of bimodule maps

Consider a unital C*-algebra A, a von Neumann algebra M, a unital sub-C*-algebra C of A and a unital *-homomorphism $π$ from C to M. Let u: A --> M be a decomposable map (i.e. a linear combination of completely positive maps) which is a C-bimodule map with respect to $π$. We show that u is a linear combination of C-bimodule completely positive maps if and only if there exists a projection e in the commutant of $π(C)$ such that u is valued in eMe and $eπ(.)e$ has a completely positive extension A --> eMe.

math.OA↗

arXiv2010

Finite rank operators in Lie ideals of nest algebras

The main theorem provides a characterisation of the finite rank operators lying in a norm closed Lie ideal of a continuous nest algebra. These operators are charaterised as those finite rank operators in the nest algebra satisfying a condition determined by a left order continuous homomorphism on the nest. A crucial fact used in the proof of this theorem is the decomposability of the finite rank operators. One shows that a finite rank operator in a norm closed Lie ideal of a continuous nest algebra can be written as a finite sum of rank one operators lying in the ideal.

math.OA↗