Machine Learning Algorithm to Predict Acidemia Using Electronic Fetal Monitoring Recording Parameters

Background: Electronic fetal monitoring (EFM) is the universal method for the surveillance of fetal well-being in intrapartum. Our objective was to predict acidemia from fetal heart signal features using machine learning algorithms. Methods: A case–control 1:2 study was carried out compromising 378 infants, born in the Miguel Servet University Hospital, Spain. Neonatal acidemia was defined as pH < 7.10. Using EFM recording logistic regression, random forest and neural networks models were built to predict acidemia. Validation of models was performed by means of discrimination, calibration, and clinical utility. Results: Best performance was attained using a random forest model built with 100 trees. The discrimination ability was good, with an area under the Receiver Operating Characteristic curve (AUC) of 0.865. The calibration showed a slight overestimation of acidemia occurrence for probabilities above 0.4. The clinical utility showed that for 33% cutoff point, missing 5% of acidotic cases, 46% of unnecessary cesarean sections could be prevented. Logistic regression and neural networks showed similar discrimination ability but with worse calibration and clinical utility. Conclusions: The combination of the variables extracted from EFM recording provided a predictive model of acidemia that showed good accuracy and provides a practical tool to prevent unnecessary cesarean sections.     

PICVib: An accurate,fast, and simple procedure to investigate selected vibrational modes at high theoretical levels

A new approach Procedure for Investigating Categories of Vibrations (PICVib) for estimating vibrational frequencies of selected modes using only the structure and energy calculations at a more demanding computational level is presented and explored. The PICVib has an excellent performance at only a small fraction of the computational demand required for a complete analytical calculation. The errors are smaller than ca. 0.5% when DFT functionals are combined with high level ab initio methods. The approach is general because it can use any quantum chemical program and electronic structure method. It is very robust because it was validated for a wide range of frequency values (ca. 20–4800 cm^–1) and systems: XH₃ (D_3h) with X = B, Al, Ga, N, P, As, O, S, and Se, YH₄ (D_4h) with Y = C, Si, and Ge, conformers of RDX, S_N2 and E2 reactions, [W(dppe)₂(NNC₅H₁₀)] complex, carbon nanotubes, and hydrogen‐bonded complexes including guanine‐cytosine pair. © 2012 Wiley Periodicals, Inc.     

Mean–variance portfolio and contribution selection in stochastic pension funding

In this paper we study the problem of simultaneous minimization of risks, and maximization of the terminal value of expected funds assets in a stochastic defined benefit aggregated pension plan. The risks considered are the solvency risk, measured as the variance of the terminal fund’s level, and the contribution risk, in the form of a running cost associated to deviations from the evolution of the stochastic normal cost. The problem is formulated as a bi-objective stochastic problem of mean–variance and it is solved with dynamic programming techniques. We find the efficient frontier and we show that the optimal portfolio depends linearly on the supplementary cost of the fund, plus an additional term due to the random evolution of benefits.     

Mechanisms of Resonant Infrared Matrix-Assisted Pulsed Laser Evaporation

For the last decade, a variant of pulsed laser ablation, Resonant-Infrared Matrix-Assisted Pulsed Laser Evaporation (RIR-MAPLE), has been studied as a deposition technique for organic and polymeric materials. RIR-MAPLE minimizes photochemical damage from direct interaction with the intense laser beam by encapsulating the polymer in a high infrared-absorption solvent matrix. This review critically examines the thermally-induced ablation mechanisms resulting from irradiation of cryogenic solvent matrices by a tunable free electron laser (FEL). A semi-empirical model is used to calculate temperatures as a function of time in the focal volume and determine heating rates for different resonant modes in two model solvents, based on the thermodynamics and kinetics of the phase transitions induced in the solvent matrices. Three principal ablation mechanisms are discussed, namely normal vaporization at the surface, normal boiling, and phase explosion. Normal vaporization is a highly inefficient polymer deposition mechanism as it relies on collective collisions with evaporating solvent molecules. Diffusion length calculations for heterogeneously nucleated vapor bubbles show that normal boiling is kinetically limited. During high-power pulsed-FEL irradiation, phase explosion is shown to be the most significant contribution to polymer deposition in RIR-MAPLE. Phase explosion occurs when the target is rapidly heated (10⁸ to 10¹⁰ K/s) and the solvent matrix approaches its critical temperature. Spontaneous density stratification (spinodal decay) within the condensed metastable phase leads to rapid homogeneous nucleation of vapor bubbles. As these vapor bubbles interconnect, large pressures build up within the condensed phase, leading to target explosions and recoil-induced ejections of polymer to a near substrate. Phase explosion is a temperature (fluence) threshold-limited process, while surface evaporation can occur even at very low fluences.     

DFT study of NH3 dissociation on Si(1 1 1)-7 × 7. The role of intermolecular interactions

The adsorption of NH₃ molecule on the Si(1 1 1)-7 × 7 surface modelled with a cluster has been studied using density functional theory (DFT). The results indicate the existence of a precursor state for the non-dissociative chemisorption. The active site for the molecular chemisorption is the adatom; while the NH₃ molecule adsorbs on the Si restatom via this preadsorbed state, the adsorption on the Si adatom is produced practically without an energy barrier. The ammonia adsorption on the adatom induces an electron transfer from the dangling bond of this atom to the dangling bond of the adjacent Si restatom, hindering this site for the adsorption of a second NH₃ incoming molecule. However, this second molecule links strongly by means of two H-bonds. The dissociative chemisorption process was studied considering one and two ammonia molecules. For the dissociation of a lonely NH₃ molecule an energy barrier of ∼0.3 eV was calculated, yielding NH₂ on the adatom and H on the restatom. When two molecules are adsorbed, the NH₃-NH₃ interaction yields the weakening of a N-H bond of the ammonia molecule adsorbed closer the Si surface. As a consequence, the dissociation barrier practically disappears. Thus, the presence of a second NH₃ molecule at the adatom-restatom pair of the Si(1 1 1)-7 × 7 surface makes the dissociative reaction self-assisted, the total adsorption process elapsing with a negligible activation barrier (less than 0.01 eV).     

Rogue waves in a multistable system

Clear evidence of rogue waves in a multistable system is revealed by experiments with an erbium-doped fiber laser driven by harmonic pump modulation. The mechanism for the rogue wave formation lies in the interplay of stochastic processes with multistable deterministic dynamics. Low-frequency noise applied to a diode pump current induces rare jumps to coexisting subharmonic states with high-amplitude pulses perceived as rogue waves. The probability of these events depends on the noise filtered frequency and grows up when the noise amplitude increases. The probability distribution of spike amplitudes confirms the rogue wave character of the observed phenomenon. The results of numerical simulations are in good agreement with experiments.     

Hölder norm estimate for the Hilbert transform in Clifford analysis

Let Ω ? ? ⁿ be a Jordan domain with d-summable boundary Γ. The main gol of this paper is to estimate the Hölder norm of a fractal version of the Hilbert transform in the Clifford analysis context acting from Hölder spaces of Clifford algebra valued functions defined on Γ. The explicit expression for the upper bound of the norm provided here is given in terms of the Hölder exponents, the diameter of Γ and certain d-sum (d > d) of the Whitney decomposition of Ω. The result obtained is applied to standard Hilbert transform for domains with left Ahlfors-David regular surface.     

Glueball spectroscopy in strongly coupled lattice gauge theories

We study the mass spectrum up to –7 (1–) log of pure three-dimensional lattice gauge theories with action (g _P) for real irreducible and small . Besides the lowest excitationm ₀–4log, we find two nearly degenerate excited statesm ₁,m ₂ withm _i–6log (i=1, 2) and (m ₁–m ₂) at leastO().Work partially supported by CNPq (Brasil)     

A branching process model for sand avalanches

An analytically solvable model for sand avalanches of noninteracting grains of sand, based on the Chapman-Kolmogorov equations, is presented. For a single avalanche, distributions of lifetimes, sizes of overflows and avalanches, and correlation functions are calculated. Some of these are exponentials, some are power laws. Spatially homogeneous distributions of avalanches are also studied. Computer simulations of avalanches of interacting grains of sand are compared to the solutions to the Chapman-Kolmogorov equations. We find that within the range of parameters explored in the simulation, the approximation of noninteracting grains of sand is a good one.     

Homogeneous and facially homogeneous self-dual cones

We show that any self-dual come in a real finite dimensional Hilbert space is homogeneous iff it is facially homogeneous in the sense of A. Connes. We develop a spectral decomposition theory for these cones which is the analogue of the usual one for self-adjoint operators on a finite-dimensional Hilbert space.