Feature selection for classification of oscillating time series

We propose a forward sequential feature selection scheme based on k‐means clustering algorithm to derive the feature subset that classifies best the time series data base, according to the criterion of the corrected Rand index. Moreover, we investigate the effect of the standardization scheme on the feature selection and propose a standardization given by the transform to standard Gaussian distribution. Our interest in this work is in classification of oscillating dynamical systems on the basis of measures computed on time series from these systems. The features to be selected are measures of linear and non‐linear analysis of time series, such as auto‐correlation and Lyapunov exponents, as well as oscillation characteristics, such as the mean magnitude of peaks. Simulations on known oscillating deterministic and stochastic systems showed that, for repeated realizations of the same classification task, the proposed feature selection scheme selected very often the same best feature subset, giving high classification accuracy for any standardization. We found that, regardless of the standardization, the highest classification accuracy could be obtained with a small feature subset, containing most frequently an oscillating‐related feature. The same setting was applied to records of epileptic electroencephalogram signals, giving varying results and dependent on the standardization.     

Nearest neighbor estimate of conditional mutual information in feature selection

Mutual information (MI) is used in feature selection to evaluate two key-properties of optimal features, the relevance of a feature to the class variable and the redundancy of similar features. Conditional mutual information (CMI), i.e., MI of the candidate feature to the class variable conditioning on the features already selected, is a natural extension of MI but not so far applied due to estimation complications for high dimensional distributions. We propose the nearest neighbor estimate of CMI, appropriate for high-dimensional variables, and build an iterative scheme for sequential feature selection with a termination criterion, called CMINN. We show that CMINN is equivalent to feature selection MI filters, such as mRMR and MaxiMin, in the presence of solely single feature effects, and more appropriate for combined feature effects. We compare CMINN to mRMR and MaxiMin on simulated datasets involving combined effects and confirm the superiority of CMINN in selecting the correct features (indicated also by the termination criterion) and giving best classification accuracy. The application to ten benchmark databases shows that CMINN obtains the same or higher classification accuracy compared to mRMR and MaxiMin at a smaller cardinality of the selected feature subset.     

Growth of ZnO nanorods on patterned templates for efficient, large-area energy scavengers

ZnO nanostructures were grown on patterned Si substrates using a cost-efficient, low-temperature process, for their future exploitation as functional cores of nanopiezotronic applications. Different substrates and growth parameters were examined in order to determine the optimum process window that will constitute a reliable, low-cost method for large-scale ZnO nanorod production. Statistical analysis was performed to assess the size and shape distribution of the nanorods in an effort to determine the energy conversion efficiency of the resulting structures.     

Go Brown: Inner-Disciplinary Conjectures

The relationship of society to its excrement is telling at both a cultural and social level. The notion of green architecture with its emphasis on the naturally pure has so far eschewed waste as a primary generative material. Alexandros Tsamis here invokes a more inclusive notion of the environment in which by-product becomes intrinsic to any project. Copyright © 2010 John Wiley & Sons, Ltd.     

Parameterizable Real Number Models for Mixed-Signal Designs Using SystemVerilog

Nowadays, the semiconductor industry directs its attention to mixed-signal System-on-Chip (SoC) applications. Main targets are the creation of accurate and fast mixed-signal SoC designs, composed of both digital and analog components, and the reduction of time to market for this kind of integrated circuits (ICs). In order to bring a mixed-signal SoC faster to the market, higher system-level simulation speed is required, with respect to traditional modeling approaches. Real Number Modelling (RNM) could be an effective solution. In this work, a sigma-delta analog-to-digital converter (ADC), a voltage-controlled oscillator (VCO) and a digital phase-locked loop (PLL) are implemented as real number models using SystemVerilog. This paper is an extended version of work previously published by the authors. Herein, more accurate and parameterizable models were created, while their validation process is analyzed and achieved using a novel metric for accuracy estimation. The proposed models’ parameterizability enhances the usability of the models to various SoC designs. Aim of this work is to underline the RNM effectiveness provided by SystemVerilog, and exhibit a way to apply RNM for modeling and simulation of widely used analog/mixed-signal (AMS) blocks. The presented real number models were compared to Verilog-A, Verilog-AMS, and transistor-level SPICE models. All tests showed that the proposed real number models based on SystemVerilog demonstrate noteworthy improvement on simulation efficiency, with respect to previous works in the literature, preserving simultaneously sufficient accuracy.

     

Electronic Structure of C60, CuPc, and C60/CuPc Nanoparticles and their Layers

The idea to form C₆₀/CuPc dispersed nanoheterojunctions by photoexcitation of a mixture of C₆₀ and CuPc nanoparticles has been realized. The electronic structure of the nanoparticles and dispersed nanoheterojunctions formed in the mixture has been characterized by UV-Vis spectroscopy and the comparing with known experimental ultraviolet photoelectron spectra and theoretical models of electronic structure of these molecules. For the mixture of C₆₀ and CuPc nanoparticles in toluene and their coating layer on the quartz substrate the band offsets of the edges of CuPc VB and lowest unoccupied molecular orbital (LUMO) of C₆₀ band are ΔE=1.55 eV and 1.4 eV, respectively. These results show clearly the presence of C₆₀/CuPc dispersed nanoheterojunctions in the solution and on the quartz surface.     

Charge transport mechanisms and memory effects in amorphous TaNx thin films

Amorphous semiconducting materials have unique electrical properties that may be beneficial in nanoelectronics, such as low leakage current, charge memory effects, and hysteresis functionality. However, electrical characteristics between different or neighboring regions in the same amorphous nanostructure may differ greatly. In this work, the bulk and surface local charge carrier transport properties of a-TaN_x amorphous thin films deposited in two different substrates are investigated by conductive atomic force microscopy. The nitride films are grown either on Au (100) or Si [100] substrates by pulsed laser deposition at 157 nm in nitrogen environment. For the a-TaN_x films deposited on Au, it is found that they display a negligible leakage current until a high bias voltage is reached. On the contrary, a much lower threshold voltage for the leakage current and a lower total resistance is observed for the a-TaN_x film deposited on the Si substrate. Furthermore, I-V characteristics of the a-TaN_x film deposited on Au show significant hysteresis effects for both polarities of bias voltage, while for the film deposited on Si hysteresis, effects appear only for positive bias voltage, suggesting that with the usage of the appropriate substrate, the a-TaN_x nanodomains may have potential use as charge memory devices.     

Charge-coupled device spectrograph for direct solar irradiance and sky radiance measurements

The characterization of a charged-coupled device (CCD) spectrograph developed at the Laboratory of Atmospheric Physics, Thessaloniki is presented. The absolute sensitivity of the instrument for direct irradiance and sky radiance measurements was determined, respectively, with an uncertainty of 4.4% and 6.6% in the UV-B, and 3% and 6% in the UV-A, visible and near-infrared (NIR) wavelength ranges. The overall uncertainty associated with the direct irradiance and the sky radiance measurements is, respectively, of the order of 5% and 7% in the UV-B, increasing to 10% for low signals [e.g., at solar zenith angles (SZAs) larger than 70 degrees ], and 4% and 6% in the UV-A, visible, and NIR. Direct solar spectral irradiance measurements from an independently calibrated spectroradiometer (Bentham DTM 300) were compared with the corresponding CCD measurements. Their agreement in the wavelength range of 310-500nm is within 0.5% +/- 1.1% (for SZA between 20 degrees and 70 degrees ). Aerosol optical depth (AOD) derived by the two instruments using direct Sun spectra and by a collocated Cimel sunphotometer [Aerosol Robotic network (AERONET)] agree to within 0.02 +/- 0.02 in the range of 315-870 nm. Significant correlation coefficients with a maximum of 0.99 in the range of 340-360 nm and a minimum of 0.90 at 870 nm were found between synchronous AOD measurements with the Bentham and the Cimel instruments.     

Atomic scale simulations of donor–vacancy pairs in germanium

Atomic scale calculations have been used to predict the structures and relative energies of defect clusters formed between n-type dopants and lattice vacancies in germanium (Ge). These include phosphorous–vacancy (P–V), arsenic–vacancy (As–V) and antimony–vacancy (Sb–V) pairs as well as larger clusters. The calculations used a plane-wave basis set in conjunction with pseudo-potentials within the generalized gradient approximation (GGA) of density functional theory (DFT). Equivalent defects in silicon (Si) are also predicted, and these demonstrate the excellent correlation of the present simulations to previous experimental and theoretical studies. The calculations highlight similarities and important differences in the formation and binding energies of clusters in Ge compared to Si. Interestingly the binding energies of the donor–vacancy pairs do not scale with dopant size.