Feature selection for k-means clustering stability: theoretical analysis and an algorithm

Stability of a learning algorithm with respect to small input perturbations is an important property, as it implies that the derived models are robust with respect to the presence of noisy features and/or data sample fluctuations. The qualitative nature of the stability property enhardens the development of practical, stability optimizing, data mining algorithms as several issues naturally arise, such as: how “much” stability is enough, or how can stability be effectively associated with intrinsic data properties. In the context of this work we take into account these issues and explore the effect of stability maximization in the continuous (PCA-based) k-means clustering problem. Our analysis is based on both mathematical optimization and statistical arguments that complement each other and allow for the solid interpretation of the algorithm’s stability properties. Interestingly, we derive that stability maximization naturally introduces a tradeoff between cluster separation and variance, leading to the selection of features that have a high cluster separation index that is not artificially inflated by the features variance. The proposed algorithmic setup is based on a Sparse PCA approach, that selects the features that maximize stability in a greedy fashion. In our study, we also analyze several properties of Sparse PCA relevant to stability that promote Sparse PCA as a viable feature selection mechanism for clustering. The practical relevance of the proposed method is demonstrated in the context of cancer research, where we consider the problem of detecting potential tumor biomarkers using microarray gene expression data. The application of our method to a leukemia dataset shows that the tradeoff between cluster separation and variance leads to the selection of features corresponding to important biomarker genes. Some of them have relative low variance and are not detected without the direct optimization of stability in Sparse PCA based k-means. Apart from the qualitative evaluation, we have also verified our approach as a feature selection method for $k$ -means clustering using four cancer research datasets. The quantitative empirical results illustrate the practical utility of our framework as a feature selection mechanism for clustering.     

Enabling user interface developers to experience accessibility limitations through visual,hearing, physical and cognitive impairment simulation

His paper presents a tool facilitating developers of user interfaces (UIs) to experience accessibility limitations that can be posed from various disabilities during the interaction of impaired users with their developments. In this respect, various aspects of visual, hearing, physical and cognitive impairments have been modelled through filters providing approximate, yet, realistic simulations over them. These filters have formed the basis for the developed tool, which can be used either on its own (as a standalone application), or be embedded in the NetBeans Integrated Development Environment. The tool, named DIAS, allows for impairment simulations to be performed over Java, mobile and web applications. Moreover, it integrates two of the most common assistive technologies (ATs), namely a screen reader and a magnifier. As a result, developers of UIs can not only experience how interaction would be affected from various impairments, but they can also understand how their developments would be perceived by impaired users through an AT. This work aims to provide an integrated, practical solution for impairment simulation, which could be easily adopted by developers, thus realistically increasing the possibilities for the future development of interactive applications that are more accessible to users with disabilities.     

Epsilon Flock: a model migration language

Model-driven engineering introduces additional challenges for controlling and managing software evolution. Today, tools exist for generating model editors and for managing models with transformation, validation, merging and weaving. There is limited support, however, for model migration—a development activity in which instance models are updated in response to metamodel evolution. In this paper, we propose conservative copy—a style of model transformation that we believe is well-suited to model migration—and Epsilon Flock—a compact model-to-model transformation language tailored for model migration. The proposed structures are evaluated by comparing the conciseness of model migration strategies written in different styles of transformation language, using several examples of evolution taken from UML and the graphical modelling framework.     

Exposing mobile malware from the inside (or what is your mobile app really doing?)

It is without a doubt that malware especially designed for modern mobile platforms is rapidly becoming a serious threat. The problem is further multiplexed by the growing convergence of wired, wireless and cellular networks, since virus writers can now develop sophisticated malicious software that is able to migrate across network domains. This is done in an effort to exploit vulnerabilities and services specific to each network. So far, research in dealing with this risk has concentrated on the Android platform and mainly considered static solutions rather than dynamic ones. Compelled by this fact, in this paper, we contribute a fully-fledged tool able to dynamically analyze any iOS software in terms of method invocation (i.e., which API methods the application invokes and under what order), and produce exploitable results that can be used to manually or automatically trace software’s behavior to decide if it contains malicious code or not. By employing real life malware we assessed our tool both manually, as well as, via heuristic techniques and the results we obtained seem highly accurate in detecting malicious code.     

Introducing web service accessibility assessment techniques through a unified quality of service context

This paper introduces novel web service (WS) accessibility assessment techniques through a unified Quality of Services (QoS) context. The goal is to enable future QoS-aware service selection systems to select and provide accessible web services, ones that are properly designed so as to allow their consumption from end-user applications, used from people with disabilities. In this line, a WS accessibility assessment Framework (WSaaF) has been developed, on the basis of WS accessibility guidelines, dealing with accessibility issues that can appear both on the presentation level of content delivered through WSs and on the content level itself. The WSaaF and its guidelines follow the rationale behind W3C WCAG 2.0-based accessibility standardization of web content. It provides the basis toward building future accessible WSs, a task that can be further facilitated by the use of an appropriate Tool (WSaaT), developed with the aim to provide automatic assessment of services, against guidelines of the proposed framework. Then, the WS accessibility attribute is introduced, as a metric that can be used in conjunction to ones typically utilized so far, within QoS-aware service selection systems. As a result, a novel unified QoS framework is proposed, incorporating the notion of accessibility in the service selection process. The proposed unified QoS framework can eventually lead to the provision of services, which are selected from appropriate repositories and better suit the special needs of people with disabilities.     

Cornet‐Like Phosphotriazine/Diamine Polymers as Reductant and Matrix for the Synthesis of Silver Nanocomposites with Antimicrobial Activity

The synthesis of silver nanoparticles attached on the surface of a hollow cornet‐like polymer matrix which served as a reductant and host matrix is described. This hybrid organic/inorganic macromolecular matrix is exhibiting anion‐exchange properties, porous structure and hollow morphologies, and absorptions in the visible light region. Due to the anion‐exchange property and the 3D orientation of the macromolecular chains the material is defining a new functional organic/inorganic hybrid. For the synthesis of nanoparticles, no other reducing agents were used and silver nanoparticles with a mean diameter of less than 20 nm were attached on the surface of the polymer, thus inheriting the composite with high antibacterial activity tested in bacterial strains and yeasts.

     

Enhancement of Bioactive Phenols and Quality Values of Olive Oil by Recycling Olive Mill Waste Water

Mature ‘Chondrolia Chalkidikis’ olives were processed in an industrial olive oil mill equipped with a three‐phase decanter. Water was added to the decanter at a 1:2 water‐to‐paste ratio. Olive mill waste water (ΟΜWW) was used to replace the added water at a rate of 50 or 100%. Following the final separation, the obtained oil was used for chemical analysis and sensory evaluation. All oils had similar acidity, peroxide and Κ values. OMWW‐treated olive oils presented higher total phenolic content and higher antioxidant activity based on DPPH and oven tests, but lower chlorophyll and carotenoids content. However, there was no significant difference between the 50 and 100% replacement. The phenolic profile of the treated olive oils analyzed by quantitative ¹Η NMR revealed more than twofold oleocanthal and oleacein as well as oleuropein and ligstroside aglycone contents than in the control. Sensory evaluation of treated oils also showed an enhancement of fruity, bitter and pungent attributes compared to the control.     

Rotordynamic Analysis of a Shaft Using Magnetorheological and Nanomagnetorheological Fluid Journal Bearings

The rotordynamic behavior of a system supported by journal bearings is critical to its reliability. A suitable method of control of the orbital motion of a shaft in a journal bearing is the use of smart lubricants, in effect fluids with controllable physical properties. There are various categories of smart lubricants. One class of smart lubricants, magnetorheological fluids, are produced as a dispersion of magnetic particles in a carrier fluid, which is usually a conventional lubricant. These particles form chains under the influence of a magnetic field, which hinder the lubricant flow, thus changing its apparent viscosity. Magnetorheological fluids (MRFs) exhibit high yield stress, low delay of response, and relatively low friction while not in their active state. A subcategory of MRFs, nanomagnetorheological fluids (NMRFs) with particle size on the nanometer scale, exhibits lower yield stress than MRFs but display high viscosity. The effect of the MRFs and NMRFs on the rotordynamic behavior of a shaft is calculated through a combined finite element and computational fluid dynamics analysis. Though the MRF with the specific geometrical configuration of the bearing is not sufficiently activated and therefore does not improve the performance of the magnetorheological journal bearing, the NMRF has the ability of limiting up to 82% the amplitude of the vibrations of the shaft.     

Minor actinide transmutation in fast reactor metal fuels irradiated for 120 and 360 equivalent full-power days

An irradiation experiment on uranium–plutonium–zirconium (U–Pu–Zr) alloys containing 5 wt% or less minor actinides (MAs) and rare earths was carried out in the Phénix fast reactor. The isotope compositions of the fuel alloys irradiated for 120 and 360 equivalent full-power days (EFPDs) were chemically analyzed by inductively coupled plasma–mass spectrometry after 3.3–5.3 years of cooling. The results of chemical analysis indicated that the discharged burnups of the fuel alloys irradiated for 120 and 360 EFPDs were 2.1–2.5 and 5.3–6.4 at%, respectively. The changes in the isotopic abundances of plutonium, americium, and curium during the irradiation experiment were assessed to discuss the transmutation performance of MA nuclides added to U–Pu–Zr alloy fuel. Multigroup three-dimensional diffusion and burnup calculations accurately predicted the changes in these isotopic abundances after fuel fabrication. An evaluation of the MA transmutation ratio based on the results of chemical analysis revealed that the quantity of MA elements in the U–19Pu–10Zr–5MA (wt%) alloy decreased by about 20% during the irradiation experiment for 360 EFPDs.     

Generalized Hirsch h-index for disclosing latent facts in citation networks

What is the value of a scientist and its impact upon the scientific thinking? How can we measure the prestige of a journal or a conference? The evaluation of the scientific work of a scientist and the estimation of the quality of a journal or conference has long attracted significant interest, due to the benefits by obtaining an unbiased and fair criterion. Although it appears to be simple, defining a quality metric is not an easy task. To overcome the disadvantages of the present metrics used for ranking scientists and journals, J. E. Hirsch proposed a pioneering metric, the now famous h-index. In this article we demonstrate several inefficiencies of this index and develop a pair of generalizations and effective variants of it to deal with scientist ranking and publication forum ranking. The new citation indices are able to disclose trendsetters in scientific research, as well as researchers that constantly shape their field with their influential work, no matter how old they are. We exhibit the effectiveness and the benefits of the new indices to unfold the full potential of the h-index, with extensive experimental results obtained from the DBLP, a widely known on-line digital library.