Learning Structure from Data and Its Application to Ozone Prediction

In this paper we propose an algorithm for structure learning in predictive expert systems based on a probabilistic network representation. The idea is to have the simplest structure (minimum number of links) with acceptable predictive capability. The algorithm starts by building a tree structure based on measuring mutual information between pairs of variables, and then it adds links as necessary to obtain certain predictive performance. We have applied this method for ozone prediction in México City, where the ozone level is used as a global indicator for the air quality in different parts of the city. It is important to predict the ozone level a day, or at least several hours in advance, to reduce the health hazards and industrial losses that occur when the ozone reaches emergency levels. We obtained as a first approximation a tree-structured dependency model for predicting ozone in one part of the city. We observe that even with only three parameters, its estimations are acceptable.A causal network representation and the structure learning techniques produced some very interesting results for the ozone prediction problem. Firstly, we got some insight into the dependence structure of the phenomena. Secondly, we got an indication of which are the important and not so important variables for ozone forecasting. Taking this into account, the measurement and computational costs for ozone prediction could be reduced. And thirdly, we have obtained satisfactory short term ozone predictions based on a small set of the most important parameters.     

Linear microstrip surface coil for MR imaging of the rat spinal cord at 4.7 T

A new design of RF coil based on a quasi-transverse electromagnetic field is described. The coil was developed for the acquisition of MR images of the rat spinal cord at 4.7 T. Different materials for the construction of the coil were tested, and the best results were obtained with Teflon. The design of the microstrip coil enables the investigator to change the length of the coil in a longitudinal direction and yields a relatively high signal-to-noise ratio due to the restricted field of view. Low RF field penetration depth also helps in suppressing motion artifacts generated by, e.g., breathing or heartbeats.     

Dielectric response and structure of amorphous hydrogenated carbon films with nitrogen admixture

The optical properties and structure of a-C:H films were modified by addition of nitrogen into the CH₄/H₂ deposition mixture. Three films prepared in capacitively coupled rf discharge were compared: (a) hydrogenated diamond like carbon film with hydrogen content of 34% and indentation hardness of 21.7 GPa, (b) hard a-C:H:N film with nitrogen content of 13% and indentation hardness of 18.5 GPa and (c) soft a-C:H:N film with nitrogen content of 10% and indentation hardness of 6.7 GPa. It is shown how the parametrized density of states model describing dielectric response of electronic interband transitions can be applied to modified a-C:H:N and how it can be combined with correct treatment of transmittance measured in infrared range using additional Gaussian peaks in joint density of phonon states. This analysis resulted in determination of film dielectric function in wide spectral range (0.045-30 eV) and provided also information about the density of states of valence and conduction bands and lattice vibrations.     

XRD and IR analysis of nanocrystalline Ni-Zn ferrite synthesized by the sol-gel method

Nanocrystalline Ni_1−xZn_xFe₂O₄ (x = 0.0, 0.2, 0.4, 0.6, 0.8, and 1.0) ferrites were synthesized by the sol-gel method. The obtained powder was annealed at temperatures from 500 to 1000 °C. X-ray diffraction (XRD) and infrared spectroscopy (IR) were used to study the structural changes taking place with varying Zn content x. The grain size decreased while the lattice parameter increases with the increase of the Zn content.     

Probing the electromagnetic field distribution within a metallic nanodisk

A Co nanolayer is used as a local probe to evaluate the vertical inhomogeneous distribution of the electromagnetic (EM) field within a resonant metallic nanodisk. Taking advantage of the direct relation between the magneto-optical activity and the electromagnetic field intensity in the Co layer, it is shown that the nonuniform EM distribution within the nanodisk under plasmon resonant conditions has maximum values close to the upper and lower flat faces, and a minimum value in the middle.     

Pixel selection for near-infrared chemical imaging (NIR-CI) discrimination between fish and terrestrial animal species in animal protein by-product meals

This paper proposes a method based on near-infrared hyperspectral imaging for discriminating between terrestrial and fish species in animal protein by-products used in livestock feed. Four algorithms (Mahalanobis distance, Kennard-Stone, spatial interpolation, and binning) were compared in order to select an appropriate subset of pixels for further partial least squares discriminant analysis (PLS-DA). The method was applied to a set of 50 terrestrial and 40 fish meals analyzed in the 1000-1700 nm range. Models were then tested using an external validation set comprising 45 samples (25 fish and 20 terrestrial). The PLS-DA models obtained using the four subset-selection algorithms yielded a classification accuracy of 99.80%, 99.79%, 99.85%, and 99.61%, respectively. The results represent a first step for the analysis of mixtures of species and suggest that NIR-CI, providing valuable information on the origin of animal components in processed animal proteins, is a promising method that could be used as part of the EU feed control program aimed at eradicating and preventing bovine spongiform encephalopathy (BSE) and related diseases.     

Semiparametric additive models under symmetric distributions

In this paper we discuss estimation and diagnostic procedures in semiparametric additive models with symmetric errors in order to permit distributions with heavier and lighter tails than the normal ones, such as Student-t, Pearson VII, power exponential, logistics I and II, and contaminated normal, among others. Such models belong to the general class of statistical models GAMLSS proposed by Rigby and Stasinopoulos (Appl. Stat. 54:507–554, 2005). A back-fitting algorithm to attain the maximum penalized likelihood estimates (MPLEs) by using natural cubic smoothing splines is presented. In particular, the score functions and Fisher information matrices for the parameters of interest are expressed in a similar notation of that used in parametric symmetric models. Sufficient conditions on the existence of the MPLEs are presented as well as some inferential results and discussions on degrees of freedom and smoothing parameter estimation. Diagnostic quantities such as leverage, standardized residual and normal curvatures of local influence under two perturbation schemes are derived. A real data set previously analyzed under normal linear models is reanalyzed under semiparametric additive models with symmetric errors.     

Multimode polarization degree of mixture optical states in integrated directional couplers

Mixture multimode optical field classical states propagating in N?×?N integrated directional couplers are analyzed by using the density matrix formalism in a N-dimensional optical space. These mutimode optical fields present a kind of generalized polarization and accordingly a definition of a multimode polarization degree is proposed. It is based on the distance measure between a mixture state and an unpolarized state in a N-dimensional optical space so that in the case N=2 the standard polarization degree is recovered. It is shown that directional couplers can reduce or increase remarkably the multimode polarization degree of a mixture state. Likewise a simple measurement technique, based on Y junctions, of this multimode polarization degree is proposed. Finally all the results can be formally extended to the special case of multimode single photon quantum states.     

Characterization and Physical Explanation of Energetic Particles on Planck HFI Instrument

The Planck High Frequency Instrument (HFI) has been surveying the sky continuously from the second Lagrangian point (L2) between August 2009 and January 2012. It operates with 52 high impedance bolometers cooled at 100 mK in a range of frequency between 100 GHz and 1 THz with unprecedented sensitivity, but strong coupling with cosmic radiation. At L2, the particle flux is about 5 \(\hbox {cm}^{-2}\,\hbox {s}^{-1}\) and is dominated by protons incident on the spacecraft. Protons with an energy above 40 MeV can penetrate the focal plane unit box causing two different effects: glitches in the raw data from direct interaction of cosmic rays with detectors (producing a data loss of about 15 % at the end of the mission) and thermal drifts in the bolometer plate at 100 mK adding non-Gaussian noise at frequencies below 0.1 Hz. The HFI consortium has made strong efforts in order to correct for this effect on the time ordered data and final Planck maps. This work intends to give a view of the physical explanation of the glitches observed in the HFI instrument in-flight. To reach this goal, we performed several ground-based experiments using protons and \(\alpha \) particles to test the impact of particles on the HFI spare bolometers with a better control of the environmental conditions with respect to the in-flight data. We have shown that the dominant part of glitches observed in the data comes from the impact of cosmic rays in the silicon die frame supporting the micro-machined bolometric detectors propagating energy mainly by ballistic phonons and by thermal diffusion. The implications of these results for future satellite missions will be discussed.     

Strain Mapping at the Atomic Scale in Highly Mismatched Heterointerfaces

A complete characterization of dislocation network in a highly mismatched interface with high spatial resolution has been performed. The interface between InN quantum dots and a (0001) GaN substrate contains three noninteracting sets of regularly‐spaced misfit dislocations lying along <110> directions. The network has a “Star of David” form, with each star bounding a hexagonal region which is pseudomorphic. These misfit dislocations form a threading dislocation network at the island edges due to free surface forces.