We introduce a novel exploratory technique, termed biarchetype analysis, which extends archetype analysis to simultaneously identify archetypes of both observations and features. This innovative unsupervised machine learning tool aims to represent observations and features through instances of pure types, or biarchetypes, which are easily interpretable as they embody mixtures of observations and features. Furthermore, the observations and features are expressed as mixtures of the biarchetypes, which makes the structure of the data easier to understand. We propose an algorithm to solve biarchetype analysis. Although clustering is not the primary aim of this technique, biarchetype analysis is demonstrated to offer significant advantages over biclustering methods, particularly in terms of interpretability. This is attributed to biarchetypes being extreme instances, in contrast to the centroids produced by biclustering, which inherently enhances human comprehension. The application of biarchetype analysis across various machine learning challenges underscores its value.

Archetypoid analysis (ADA) is an exploratory approach that explains a set of continuous observations as mixtures of pure (extreme) patterns. Those patterns (archetypoids) are actual observations of the sample which makes the results of this technique easily interpretable, even for non-experts. Note that the observations are approximated as a convex combination of the archetypoids. Archetypoid analysis, in its current form, cannot be applied directly to ordinal data. We propose and describe a two-step method for applying ADA to ordinal responses based on the ordered stereotype model. One of the main advantages of this model is that it allows us to convert the ordinal data to numerical values, using a new data-driven spacing that better reflects the ordinal patterns of the data, and this numerical conversion then enables us to apply ADA straightforwardly. The results of the novel method are presented for two behavioural science applications. Finally, the proposed method is also compared with other unsupervised statistical learning methods.

The problem of detecting outliers in multivariate data sets with continuous numerical features is addressed by a new method. This method combines projections into relevant subspaces by archetype analysis with a nearest neighbor algorithm, through an appropriate ensemble of the results. Our method is able to detect an anomaly in a simple data set with a linear correlation of two features, while other methods fail to recognize that anomaly. Our method performs among top in an extensive comparison with 23 state-of-the-art outlier detection algorithms with several benchmark data sets. Finally, a novel industrial data set is introduced, and an outlier analysis is carried out to improve the fit of footwear, since this kind of analysis has never been fully exploited in the anthropometric field.

The present work describes the 100% virtual ATIDES (Avances en Tecnologı ´as, Innovacio ´n y Desafı ´os de la Educacio ´n Superior) conference that was held between October 15 and 31, 2018, sponsored by Universitat Jaume I (UJI), Spain.Online conferences like this have been the subject of some controversy in the field of education over the last decade.Indeed, we have found a few texts that are against them.One of these is [1], whose authors claim that "interaction is not enough" to ensure efficient simulation of face-to-face contact.However, the Canadian academic community (for instance, the Centre for Distance Education at Athabasca University) is a strong advocate of online conferences (see [2,3]).Among other advantages, this kind of conference is "family-friendly," i.e., they break barriers for researchers with family obligations [4], in particular many women [5].In addition, these conferences overcome the drawback of parallel sessions at face-to-face conferences, at which participants must choose certain talks and miss others.Anderson and Anderson [6] even put forward environmental and economic arguments: "Transportation is a major contributor of carbon dioxide (CO2) emissions."On the other hand, Abdullah [7] and Kear, Chetwynd, and Jefferis [8] look at the matter from another point of view that is also important: Social presence at online conferences.Gichora and colleagues [9] propose 10 rules for organizing a virtual conference, but they only focus on real-time virtual conferences.Real-time virtual conferences emulate the structure of well-known on-site conferences, with the use of video-conferencing tools for connecting the participants in a virtual meeting.We, on the other hand, focus on non-real-time web conferences.In this kind of conference, a web platform connects the participants, but the event does not take place live.Communication can be made through posts on a forum, which can sometimes take hours or even days.Non-real-time web conferences share many of the advantages of real-time virtual conferences, such as being family friendly and allowing the involvement of participants with a low budget for traveling, which makes this kind of conference more participative and inclusive.However, non-real-time web conferences have other advantages that real-time web conferences do not.Asynchronous interaction allows the attendees to have more time for reflection before they put forward their questions and comments, and, likewise, the speaker has enough time to prepare an appropriate answer.This makes up for the lack of face-to-face social interaction in this type of conference.In fact, numerous interactions occur via forums and are, in many cases, numerous and more informal and flexible than at classical scientific meetings [10].Forums give rise to stimulating discussions like those that can occur at unconferences [11].Furthermore, this asynchrony allows participants from all over the world to take part without worrying about time zone differences.

BACKGROUND: Random forests are a popular method in many fields since they can be successfully applied to complex data, with a small sample size, complex interactions and correlations, mixed type predictors, etc. Furthermore, they provide variable importance measures that aid qualitative interpretation and also the selection of relevant predictors. However, most of these measures rely on the choice of a performance measure. But measures of prediction performance are not unique or there is not even a clear definition, as in the case of multivariate response random forests. METHODS: A new alternative importance measure, called Intervention in Prediction Measure, is investigated. It depends on the structure of the trees, without depending on performance measures. It is compared with other well-known variable importance measures in different contexts, such as a classification problem with variables of different types, another classification problem with correlated predictor variables, and problems with multivariate responses and predictors of different types. RESULTS: Several simulation studies are carried out, showing the new measure to be very competitive. In addition, it is applied in two well-known bioinformatics applications previously used in other papers. Improvements in performance are also provided for these applications by the use of this new measure. CONCLUSIONS: This new measure is expressed as a percentage, which makes it attractive in terms of interpretability. It can be used with new observations. It can be defined globally, for each class (in a classification problem) and case-wise. It can easily be computed for any kind of response, including multivariate responses. Furthermore, it can be used with any algorithm employed to grow each individual tree. It can be used in place of (or in addition to) other variable importance measures.

I propose the use of h‐plots for visualizing the asymmetric relationships between the citing and cited profiles of journals in a common map. With this exploratory tool, we can understand better the journal's dual roles of citing and being cited in a reference network. The h‐plot is introduced and its use is validated with a set of 25 journals belonging to the statistics area. The relatedness factor is considered for describing the relations of citations from a journal “i” to a journal “j,” and the citations from the journal “j” to the journal “i.” More information has been extracted from the h‐plot, compared with other statistical techniques for modelling and representing asymmetric data, such as multidimensional unfolding.

The goal of this paper is to segment high-resolution images of natural landscapes into different cover types. With this aim, morphological texture features (descriptors of random sets obtained by morphological transformations) are used in order to avoid the limitations of spectral features. First, a supervised segmentation (the textures to detect having been previously determined) is presented. The classes correspond to different degrees of tree densities. Second, a methodology for an unsupervised texture segmentation (no a priori information about the textures is supplied) is proposed. The number of classes is automatically determined. The proposed procedures have been tested on several images, providing promising results

Image compression systems commonly operate by transforming the input signal into a new representation whose elements are independently quantized. The success of such a system depends on two properties of the representation. First, the coding rate is minimized only if the elements of the representation are statistically independent. Second, the perceived coding distortion is minimized only if the errors in a reconstructed image arising from quantization of the different elements of the representation are perceptually independent. We argue that linear transforms cannot achieve either of these goals and propose, instead, an adaptive nonlinear image representation in which each coefficient of a linear transform is divided by a weighted sum of coefficient amplitudes in a generalized neighborhood. We then show that the divisive operation greatly reduces both the statistical and the perceptual redundancy amongst representation elements. We develop an efficient method of inverting this transformation, and we demonstrate through simulations that the dual reduction in dependency can greatly improve the visual quality of compressed images.

The goal of this work is to segment high resolution images of natural landscapes into different cover types. With this aim, morphological texture descriptors are used in order to avoid the limitations of spectral features. Firstly, an unsupervised texture segmentation method is presented, since no a priori information about the textures is supplied. Secondly, the same features are utilized in a supervised case, where the textures to detect have been previously determined.

Texture classification of an image or subimage is an important problem in texture analysis. Many procedures have been proposed. A global framework for texture classification based on random closed set theory is proposed in this paper. In this approach, a binary texture is considered as an outcome of a random closed set. Some distributional descriptors of this stochastic model are used as texture features in order to classify the binary texture, in particular spherical and linear contact distributions and K-functions. If a grayscale texture has to be classified, then the original texture is reduced to a multivariate random closed set where each component (a different random set) corresponds with those pixels verifying a local property. Again, some functional descriptors of the multivariate random closed set defined from the texture can be used as texture features to describe and classify the grayscale texture. Marginal and cross spherical and linear contact distributions and K-functions have been used. Experimental validation is provided by using Brodatz's database and another standard texture database.

In this paper, a multigrid motion compensation video coder based on the current human visual system (HVS) contrast discrimination models is proposed. A novel procedure for the encoding of the prediction errors has been used. This procedure restricts the maximum perceptual distortion in each transform coefficient. This subjective redundancy removal procedure includes the amplitude nonlinearities and some temporal features of human perception. A perceptually weighted control of the adaptive motion estimation algorithm has also been derived from this model. Perceptual feedback in motion estimation ensures a perceptual balance between the motion estimation effort and the redundancy removal process. The results show that this feedback induces a scale-dependent refinement strategy that gives rise to more robust and meaningful motion estimation, which may facilitate higher level sequence interpretation. Perceptually meaningful distortion measures and the reconstructed frames show the subjective improvements of the proposed scheme versus an H.263 scheme with unweighted motion estimation and MPEG-like quantization.

Fovea segmentation in fluorescein angiographies is a fundamental first task in any study of ocular diseases. The importance of fovea detection is due to the fact that the nearer the centre of the fovea a lesion is, the graver this lesion is. The proposed method is based on B-snakes and uses a greedy algorithm to minimise an appropriate energy which accurately leads to a convenient characterisation of the boundary of the foveal zone. The first initialisation step, which consists of finding the most appropriate local minimum along with a procedure to construct an initial contour involving a region growing algorithm, leads to a convenient and robust initialisation of the proposed active contour model.