Learning from time series:Supervised Aggregative Feature Extraction

Many modeling problems require to estimate a scalar output from one or more time series. Such problems are usually tackled by extracting a fixed number of features from the time series (like their statistical moments), with a consequent loss in information that leads to suboptimal predictive models. Moreover, feature extraction techniques usually make assumptions that are not met by real world settings (e.g. uniformly sampled time series of constant length), and fail to deliver a thorough methodology to deal with noisy data. In this paper a methodology based on functional learning is proposed to overcome the aforementioned problems; the proposed Supervised Aggregative Feature Extraction (SAFE) approach allows to derive continuous, smooth estimates of time series data (yielding aggregate local information), while simultaneously estimating a continuous shape function yielding optimal predictions. The SAFE paradigm enjoys several properties like closed form solution, incorporation of first and second order derivative information into the regressor matrix, interpretability of the generated functional predictor and the possibility to exploit Reproducing Kernel Hilbert Spaces setting to yield nonlinear predictive models. Simulation studies are provided to highlight the strengths of the new methodology w.r.t. standard unsupervised feature selection approaches. © 2012 IEEE.

CALCULATION OF SLOW MODE SURFACE-PLASMON POLARITON PROPERTIES RELATED TO EXPERIMENTAL-OBSERVATIONS

Recent experimental results definitively showed, for the first time, optical radiation mediated by the slow mode surface plasmon polariton of metal-oxide-metal tunnel junctions. Here, dispersion curves for this mode are calculated. They are consistent with first-order grating coupling to light at the energies of the experimental emission peaks. The curves are then used to analyze second-order and high-energy (> 2.35 eV) grating coupling of the polaritons to radiation. Finally, variation of slow mode damping as a function of energy is used to explain qualitatively the relative experimental peak emission intensities and the absence of radiation peaks above 2.35 eV.

Arsenic(III,V) adsorption onto charred dolomite: charring optimization and batch studies

In this work, the removal of arsenic from aqueous solutions onto thermally processed dolomite is investigated. The dolomite was thermally processed (charred) at temperatures of 600, 700 and 800 degrees C for 1, 2, 4 and 8 h. Isotherm experiments were carried out on these samples over a wide pH range. A complete arsenic removal was achieved over the pH range studied when using the 800 degrees C charred dolomite. However, at this temperature, thermal degradation of the dolomite weakens its structure due to the decomposition of the magnesium carbonate, leading to a partial dissolution. For this reason, the dolomitic sorbent chosen for further investigations was the 8 h at 700 degrees C material. Isotherm studies indicated that the Langmuir model was successful in describing the process to a better extent than the Freundlich model for the As(V) adsorption on the selected charred dolomite. However, for the As(III) adsorption, the Freundlich model was more successful in describing the process. The maximum adsorption capacities of charred dolomite for arsenite and arsenate ions are 1.846 and 2.157 mg/g, respectively. It was found that both the pseudo first- and second-order kinetic models are able to describe the experimental data (R-2 > 0.980). The data suggest the charring process allows dissociation of the dolomite to calcium carbonate and magnesium oxide, which accelerates the process of arsenic oxide and arsenic carbonate precipitation. (C) 2014 Elsevier B.V. All rights reserved.

Emotions and voting in EU referendums

There is an emerging scholarship on the emotional bases of political opinion and behaviour and, in particular, the contrasting implications of two distinct negative emotions - anger and anxiety. I apply the insights in this literature to the previously unresearched realm of the emotional bases of voting in EU referendums. I hypothesise that anxious voters rely on substantive EU issues and angry voters rely on second-order factors relating to domestic politics (partisanship and satisfaction with government). Focusing on the case of Irish voting in the Fiscal Compact referendum, and using data from a representative sample of voters, I find support for the hypotheses and discuss the implications of the findings for our understanding of the emotional conditionality of EU referendum voting. 

Chemistry and Mechanism of Alizarin Red S and Methylene Blue Biosorption onto Olive Stone: Equilibrium and Kinetic

The biosorption process of anionic dye Alizarin Red S (ARS) and cationic dye methylene blue (MB) as a function of solution pH, initial concentration and contact time onto olive stone (OS) biomass has been investigated. The main objectives of the current study are to: (i) study the chemistry and the mechanism of ARS and MB biosorption onto olive stone and the type of OS–ARS, MB interactions occurring, (ii) study the biosorption equilibrium and kinetic experimental data required for the design and operation of column reactors. Equilibrium biosorption isotherms and kinetics were also examined. Experimental equilibrium data were fitted to four different isotherms by non-linear regression method, however, the biosorption experimental data for ARS and MB dyes were well interpreted by the Temkin and Langmuir isotherms, respectively. The maximum monolayer adsorption capacity for ARS and MB dyes were 109.0 and 102.6 mg/g, respectively. The kinetic data of the two dyes could be better described by the pseudo second-order model. The data showed that olive stone can be effectively used for removing dyes from wastewater.

Ultracold, radiative charge transfer in hybrid Yb ion – Rb atom traps

Ultracold hybrid ion–atom traps offer the possibility of microscopic manipulation of quantum coherences in the gas using the ion as a probe. However, inelastic processes, particularly charge transfer can be a significant process of ion loss and has been measured experimentally for the ${\rm Y}{{{\rm b}}^{+}}$ ion immersed in a Rb vapour. We use first-principles quantum chemistry codes to obtain the potential energy curves and dipole moments for the lowest-lying energy states of this complex. Calculations for the radiative decay processes cross sections and rate coefficients are presented for the total decay processes; ${\rm Y}{{{\rm b}}^{+}}(6{\rm s}{{\;}^{2}}{\rm S})+{\rm Rb}(5{\rm s}{{\;}^{2}}{\rm S})\to {\rm Yb}(6{{{\rm s}}^{2}}{{\;}^{1}}{\rm S})+{\rm R}{{{\rm b}}^{+}}(4{{{\rm p}}^{6}}{{\;}^{1}}{\rm S})+h\nu $ and ${\rm Y}{{{\rm b}}^{+}}(6{\rm s}{{\;}^{2}}{\rm S})+{\rm Rb}(5{\rm s}{{\;}^{2}}{\rm S})\to {\rm YbR}{{{\rm b}}^{+}}({{X}^{1}}{{\Sigma }^{+}})+h\nu $. Comparing the semi-classical Langevin approximation with the quantum approach, we find it provides a very good estimate of the background at higher energies. The results demonstrate that radiative decay mechanisms are important over the energy and temperature region considered. In fact, the Langevin process of ion–atom collisions dominates cold ion–atom collisions. For spin-dependent processes [1] the anisotropic magnetic dipole–dipole interaction and the second-order spin–orbit coupling can play important roles, inducing coupling between the spin and the orbital motion. They measured the spin-relaxing collision rate to be approximately five orders of magnitude higher than the charge-exchange collision rate [1]. Regarding the measured radiative charge transfer collision rate, we find that our calculation is in very good agreement with experiment and with previous calculations. Nonetheless, we find no broad resonances features that might underly a strong isotope effect. In conclusion, we find, in agreement with previous theory that the isotope anomaly observed in experiment remains an open question.

Simultaneous biosorption of methylene blue and trivalent chromium onto olive stone

In this work, olive stone (OS) was utilized to investigate its capacity as biosorbent for methylene blue (MB) and Cr(III), which are usually present in textile industry effluents. Equilibrium and kinetic experiments were performed in batch experiments. The biosorption process followed pseudo-second-order kinetics. The equilibrium data were fitted with several models, but Langmuir and Sips models best reproduced the experimental results. Maximum biosorption capacities were 3.296 mg/g (0.0116 mmol/g) and 4.990 mg/g (0.0960 mmol/g) for MB and Cr(III), respectively. Several operation variables, such as
biosorbent mass, flow rate, and initial concentration on the removal of dye and metal, were evaluated in column system. The removal efficiency improved as OS mass increased and decreased when flow rate and initial concentration increased. Also, MB uptake was substantially decreased by increasing the initial concentration of Cr(III), ranging from 6.09 to 2.75 mg/g. These results show that the presence of Cr(III) significantly modifies the biosorption capacity of MB by the OS. These results suggest that OS is a potential low-cost food industry waste for textile industry wastewater treatment.

Mechanism of Alizarin Red S and Methylene Blue Biosorption onto Olive Stone: Isotherm study in Single and Binary Systems

The biosorption process of anionic dye Alizarin Red S (ARS) and cationic dye methylene blue (MB) as a function of contact time, initial concentration and solution pH onto olive stone (OS) biomass has been investigated. Equilibrium biosorption isotherms in single and binary systems and kinetics in batch mode were also examined. The kinetic data of the two dyes were better described by the pseudo second-order model. At low concentration, ARS dye appeared to follow a two-step diffusion process, while MB dye followed a three-step diffusion process. The biosorption experimental data for ARS and MB dyes were well suited to the Redlich-Peterson isotherm. The maximum biosorption of ARS dye, qmax = 16.10 mg/g, was obtained at pH 3.28 and the maximum biosorption of MB dye, qmax = 13.20 mg/g, was observed at basic pH values. In the binary system, it was indicated that the MB dye diffuses firstly inside the biosorbent particle and occupies the biosorption sites forming a monodentate complex and then the ARS dye enters and can only bind to untaken sites; forms a tridentate complex with OS active sites.

Active site labeling of cysteine cathepsins by a straightforward diazomethylketone probe derived from the N-terminus of human cystatin C

We designed a straightforward biotinylated probe using the N-terminal substrate-like region of the inhibitory site of human cystatin C as a scaffold, linked to the thiol-specific reagent diazomethylketone group as a covalent warhead (i.e. Biot-(PEG)2-Ahx-LeuValGly-DMK). The irreversible activity-based probe bound readily to cysteine cathepsins B, L, S and K. Moreover affinity labeling is sensitive since active cathepsins were detected in the nM range using an ExtrAvidin®-peroxidase conjugate for disclosure. Biot-(PEG)2-Ahx-LeuValGly-DMK allowed a slightly more pronounced labeling for cathepsin S with a compelling second-order rate constant for association (kass = 2,320,000 M−1 s−1). Labeling of the active site is dose-dependent as observed using 6-cyclohexylamine-4-piperazinyl-1,3,5-triazine-2-carbonitrile, as competitive inhibitor of cathepsins. Finally we showed that Biot-(PEG)2-Ahx-LeuValGly-DMK may be a simple and convenient tool to label secreted and intracellular active cathepsins using a myelomonocytic cell line (THP-1 cells) as model.

Identification of Dynamics of Surface Suction Over an Airfoil at Low Reynolds Numbers

This paper deals with identification of dynamics in suction control of airfoils for low Reynolds number regimes (8 x 10^4 - 5 x 10^5). In particular, the dynamics of interest is the map that relates suction pressure and surface pressure. Identification of such dynamics is of use to a variety of active control applications including suction control in small/medium wind turbines which operate in these Reynolds number regimes. Prior research has largely focused on higher Reynolds number regimes, creating a need for such a study. Towards identifying the said dynamic relations, experiments were conducted on NACA0012 airfoil in a wind tunnel. The dynamic relation between suction and surface pressure was identified as an overdamped second order system.

Supervised Aggregative Feature Extraction for Big Data Time Series Regression

In many applications, and especially those where batch processes are involved, a target scalar output of interest is often dependent on one or more time series of data. With the exponential growth in data logging in modern industries such time series are increasingly available for statistical modeling in soft sensing applications. In order to exploit time series data for predictive modelling, it is necessary to summarise the information they contain as a set of features to use as model regressors. Typically this is done in an unsupervised fashion using simple techniques such as computing statistical moments, principal components or wavelet decompositions, often leading to significant information loss and hence suboptimal predictive models. In this paper, a functional learning paradigm is exploited in a supervised fashion to derive continuous, smooth estimates of time series data (yielding aggregated local information), while simultaneously estimating a continuous shape function yielding optimal predictions. The proposed Supervised Aggregative Feature Extraction (SAFE) methodology can be extended to support nonlinear predictive models by embedding the functional learning framework in a Reproducing Kernel Hilbert Spaces setting. SAFE has a number of attractive features including closed form solution and the ability to explicitly incorporate first and second order derivative information. Using simulation studies and a practical semiconductor manufacturing case study we highlight the strengths of the new methodology with respect to standard unsupervised feature extraction approaches.

Spatio-temporal Rich Model Based Video Steganalysis on Cross Sections of Motion Vector Planes

A rich model based motion vector steganalysis benefiting from both temporal and spatial correlations of motion vectors is proposed in this work. The proposed steganalysis method has a substantially superior detection accuracy than the previous methods, even the targeted ones. The improvement in detection accuracy lies in several novel approaches introduced in this work. Firstly, it is shown that there is a strong correlation, not only spatially but also temporally, among neighbouring motion vectors for longer distances. Therefore, temporal motion vector dependency along side the spatial dependency is utilized for rigorous motion vector steganalysis. Secondly, unlike the filters previously used, which were heuristically designed against a specific motion vector steganography, a diverse set of many filters which can capture aberrations introduced by various motion vector steganography methods is used. The variety and also the number of the filter kernels are substantially more than that of used in previous ones. Besides that, filters up to fifth order are employed whereas the previous methods use at most second order filters. As a result of these, the proposed system captures various decorrelations in a wide spatio-temporal range and provides a better cover model. The proposed method is tested against the most prominent motion vector steganalysis and steganography methods. To the best knowledge of the authors, the experiments section has the most comprehensive tests in motion vector steganalysis field including five stego and seven steganalysis methods. Test results show that the proposed method yields around 20% detection accuracy increase in low payloads and 5% in higher payloads.