University Distrital

Bogotá, Colombia

University Distrital

Bogotá, Colombia
Time filter
Source Type

Rodriguez N.,University Distrital | Rojas-Galeano S.,University Distrital
BioData Mining | Year: 2017

Background: Discovering relevant features (biomarkers) that discriminate etiologies of a disease is useful to provide biomedical researchers with candidate targets for further laboratory experimentation while saving costs; dependencies among biomarkers may suggest additional valuable information, for example, to characterize complex epistatic relationships from genetic data. The use of classifiers to guide the search for biomarkers (the so-called wrapper approach) has been widely studied. However, simultaneously searching for relevancy and dependencies among markers is a less explored ground. Results: We propose a new wrapper method that builds upon the discrimination power of a weighted kernel classifier to guide the search for a probabilistic model of simultaneous marginal and interacting effects. The feasibility of the method was evaluated in three empirical studies. The first one assessed its ability to discover complex epistatic effects on a large-scale testbed of generated human genetic problems; the method succeeded in 4 out of 5 of these problems while providing more accurate and expressive results than a baseline technique that also considers dependencies. The second study evaluated the performance of the method in benchmark classification tasks; in average the prediction accuracy was comparable to two other baseline techniques whilst finding smaller subsets of relevant features. The last study was aimed at discovering relevancy/dependency in a hepatitis dataset; in this regard, evidence recently reported in medical literature corroborated our findings. As a byproduct, the method was implemented and made freely available as a toolbox of software components deployed within an existing visual data-mining workbench. Conclusions: The mining advantages exhibited by the method come at the expense of a higher computational complexity, posing interesting algorithmic challenges regarding its applicability to large-scale datasets. Extending the probabilistic assumptions of the method to continuous distributions and higher-degree interactions is also appealing. As a final remark, we advocate broadening the use of visual graphical software tools as they enable biodata researchers to focus on experiment design, visualisation and data analysis rather than on refining their scripting programming skills. © 2017 The Author(s).

Medina J.A.V.,University Distrital | Escalante P.A.A.,University of Los Andes, Colombia | Salcedo I.A.L.,National University of Costa Rica
Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) | Year: 2017

A common strategy in high spatial resolution image analysis is to define coarser geometric space elements, i.e. superpixels, by grouping near pixels based on (a, b)–connected graphs as neighborhood definitions. Such an approach, however, cannot meet some topological axioms needed to ensure a correct representation of connectedness relationships. Superpixel boundaries may present ambiguities because one-dimensional contours are represented by pixels, which are two-dimensional. Additionally, the high spatial resolution available today has increased the volume of data that must be processed during image segmentation even after data reduction phases such as principal component analysis. The inherent complexity of segmentation algorithms, including texture analysis, along with the aforementioned volume of data, demands considerable computing resources. In this paper, we propose a novel way for segmenting hyperspectral imagery data by defining a computational framework based on Axiomatic Locally Finite Spaces (ALFS) provided by Cartesian complexes, which provide a geometric space that complies with the T0 digital topology. Our approach links also oriented matroids to geometric space representations and is implemented on a parallel computational framework. We evaluated quantitatively our approach on a subset of hyperspectral remote sensing scenes. Our results show that, by departing from the conventional pixel representation, it is possible to segment an image based on a topologically correct digital space, while simultaneously taking advantage of combinatorial features of their associated oriented matroids. © Springer International Publishing AG 2017.

Cortes-B. R.C.-B,University Distrital | Motley T.J.,Old Dominion University
Phytotaxa | Year: 2015

The Henriquezieae-Posoquerieae-Sipaneeae clade (Rubiaceae) is centered in the Guayana Region of Northern South America. The Henriquezieae and Sipaneeae contain a large number of taxa that are endemic to the Guayana Region, while the Posoquerieae exhibits a wider distribution throughout the Neotropics. In order to evaluate the phylogenetic relationships within this clade, and assess the evolution of selected morphological characters potentially important in its diversification, we performed parsimony analyses of ITS, trnL-F, rps16 and morphology, Bayesian Markov chain Monte Carlo analyses of the molecular data, and optimized six morphological characters on cpDNA trees. Results confirm the monophyly of each of the tribes, and that of the genera included in the study. The Sipaneeae are the earliest diverging lineage, sister to a well-supported Henriquezieae-Posoquerieae clade. Henriquezieae have many traits unique or atypical in Rubiaceae, a tendency observed in other lineages endemic to the Guayana Region, suggesting strong selection for morphological novelties necessary to colonize the extremely poor soils of this region. The pollen catapult mechanism originated in the Molopanthera-Posoqueria clade is perhaps unique in angiosperms, and may represent a novel strategy that played an important role in the early divergence that gave rise to both genera, and to the early radiation of the genus Posoqueria. Diversification of Posoqueria outside the Guayana Region probably was accompanied by a shift from capsular to baccate fruits that facilitated colonization of understory forests. Habit specialization may be considered a consistent factor in the diversification of the Guayanan-centered clade of the Rubiaceae. © 2015 Magnolia Press

Mejia M.,University of Los Andes, Colombia | Pena N.,University of Los Andes, Colombia | Munoz J.L.,Polytechnic University of Catalonia | Esparza O.,Polytechnic University of Catalonia | Alzate M.A.,University Distrital
Journal of Network and Computer Applications | Year: 2011

Cooperation among nodes is fundamental for the operation of mobile ad hoc networks (MANETs). In such networks, there could be selfish nodes that use resources from other nodes to send their packets but that do not offer their resources to forward packets for other nodes. Thus, a cooperation enforcement mechanism is necessary. Trust models have been proposed as mechanisms to incentive cooperation in MANETs and some of them are based on game theory concepts. Among game theoretic trust models, those that make nodes' strategies evolve genetically have shown promising results for cooperation improvement. However, current approaches propose a highly centralized genetic evolution which render them unfeasible for practical purposes in MANETs. In this article, we propose a trust model based on a non-cooperative game that uses a bacterial-like algorithm to let the nodes quickly learn the appropriate cooperation behavior. Our model is completely distributed, achieves optimal cooperation values in a small fraction of time compared with centralized algorithms, and adapts effectively to environmental changes. © 2010 Elsevier Ltd. All rights reserved.

Mejia M.,University Militar Nueva Granada | Mejia M.,University of Los Andes, Colombia | Pena N.,University of Los Andes, Colombia | Munoz J.L.,Polytechnic University of Catalonia | And 2 more authors.
Ad Hoc Networks | Year: 2012

The scarce resources of a mobile ad hoc network (MANET) should not be wasted attending selfish nodes (those nodes that use resources from other nodes to send their own packets, without offering their own resources to forward other nodes' packets). Thus, rational nodes (those nodes willing to cooperate if deemed worthy) must detect and isolate selfish nodes in order to cooperate only among themselves. To achieve this purpose, in this paper we present a new game theoretic trust model called DECADE (Distributed Emergent Cooperation through ADaptive Evolution). The design of DECADE is shown by first, analyzing a simple case of packet forwarding between two nodes, and then the results are extended to bigger networks. In DECADE, each node seeks individually to maximize its chance to deliver successfully their own packets, so that the cooperation among rational nodes and the isolation of selfish nodes appear as an emergent collective behavior. This behavior emerges as long as there is a highly dynamic interaction among nodes. So, for those cases where the mobility alone does not suffice to provide this interaction, DECADE includes a sociability parameter that encourages nodes to interact among them for faster learning and adaptability. Additionally, DECADE introduces very low overhead on computational and communication resources, achieving close to optimal cooperation levels among rational nodes and almost complete isolation of selfish nodes. © 2012 Elsevier B.V. All rights reserved.

Pinzon C.M.,University Distrital | Rojas-Galeano S.,University Distrital
2015 10th Colombian Computing Conference, 10CCC 2015 | Year: 2015

User-generated text is the primary source of interaction in virtual communities on Web2.0 applications such as forums, blogs or social networks. Unfortunately some users abuse this freedom of speech liberty to disseminate non-authorised profanity content (foul language, insults, advertisement, boosting or denigration of a name or a trademark). Naïve filters based on literal comparisons against black-lists of forbidden terms, fail to detect variations obtained by character transliteration or masking (e.g. writing piss as P!55 or p.i.s.s). Recent approaches to this problem inspired in sequence alignment methods from comparative genomics in bioinformatics, have shown promise in preventing overlooking such variants. Building upon those results we have developed an experimental Web forum allowing users to generate text that is screened against transliterated profanity. In this paper we introduce the software (ForumForte) and describe briefly the technique and engineering behind it. We anticipate this kind of tools might prove beneficial for content moderation in mainstream applications such as newspaper forums and micro-blogging social networking sites. Our software is open-source under the New BSD License and is available at: © 2015 IEEE.

Unsolicited bulk email (spam) nowadays accounts for nearly 75% of daily email traffic, a figure that speaks strongly for the need of finding better protection mechanisms against its dissemination. A clever trick recently exploited by email spammers in order to circumvent textual-based filters, involves obfuscation of black-listed words with visually equivalent text substitutions from non-alphabetic symbols, in such a way it still conveys the semantics of the original word to the human eye (e.g. masking viagra as v1@gr@ or as v-i-a-g-r-a). In this paper we discuss how a simple-yet-effective adaptation of a classical algorithm for string matching may meet this stylish challenge to effectively reveal the similarity between genuine spam-trigger terms with their disguised alpha-numeric variants.

Cortes R.,University Distrital
Novon | Year: 2014

Retiniphyllum francoanum Cortés-Ballén (Rubiaceae), a species restricted to the westernmost limit of the Guayana Shield in Colombia, is described. It differs from the other known species of the genus by the ovaries with seven or eight locules, and from its closest relative, R. secundiflorum Bonpl., by the shorter inflorescences, bracteole shape, colleters inside bracteoles, longer corolla tube, turbinate ovary, and by the winged and smaller pyrenes.

Barrera M.A.R.,University Distrital | Carpes W.P.,Jr.
COMPEL - The International Journal for Computation and Mathematics in Electrical and Electronic Engineering | Year: 2016

Walter Pereira Carpes Jr,Purpose - The purpose of this paper is to present the results of a particle swarm optimization (PSO) method applied in the design of a square-loop frequency selective surface (FSS) via the equivalent circuit model (ECM), considering the dielectric effective permittivity as a variable in the optimization problem. Design/methodology/approach - In the optimization process considered, besides the FSS square loop geometric parameters, the thickness and relative permittivity of dielectric material used as support are included as variables in the search space, using for this a model of dielectric effective permittivity introduced by the authors in a previous work. Findings - Square loops were designed and the obtained results were compared with designs reported in literature for applications in wireless local area network and long-Term evolution 4G systems. The low computational cost is remarkable as well as the acceptable accuracy obtained with the proposed approach. The PSO method results were implemented with the ECM and compared with those obtained via Ansys - high frequency structure simulator commercial software simulations. Originality/value - The lack of a model of dielectric effective permittivity for the ECM causes a restricted search space in the stochastic FSS design process limited to only geometric parameters, as it is reported in the available literature. The proposed approach simplifies and makes more flexible the design process, and allows guiding the FSS design to unit cell surface and/or dielectric thickness of small dimensions. © Emerald Group Publishing Limited.

Rojas H.,National University of Colombia | Cortes C.,National University of Colombia | Santamaria F.,University Distrital
2012 31st International Conference on Lightning Protection, ICLP 2012 | Year: 2012

In recent years, several activities have been carried out by EMC-UNC (research group) to measure, study and classify the parameters of lightning electric fields signals in Bogotá, Colombia. However, measurements are distorted and exhibit significant noise levels caused by undesired signals from the electromagnetic environment together with other types of disturbances produced by the measuring system, which inevitably introduces noise in signal measurements. This paper shows the results of applying the fractional Fourier transform (FRFT) as an alternative technique to reduce the presence of noise and also of undesired signal components within a set of electric field signatures. Basic conditions to carry out the filtering process in the fractional Fourier domain (FRFd) are presented. Afterwards, some lightning electric field temporal characteristics are analyzed by using the processed signatures. Important differences between processed and measured waveforms are observed. © 2012 IEEE.

Loading University Distrital collaborators
Loading University Distrital collaborators