一个新的指数削减椭圆形分布族

A new family of univariate exponential slash distribution is introduced, which is based on elliptical distributions and defined by means of a stochastic representation as the scale mixture of an elliptically distributed random variable with respect to the power of an exponential random variable. The same idea is extended to the multivariate case. General properties of the resulting families, including their moments and kurtosis coefficient, are studied. And inferences based on methods of moment and maximum likelihood are discussed. A real data is presented to show this family is flexible and fits much better than other related families.     

Elastomeric Properties of Polysiloxane Networks. Bimodal Elastomers that are Spatially Inhomogeneous and Others that are Very Broadly Multimodal

End‐linking poly(dimethylsiloxane) was used to prepare bimodal elastomers networks so as to have inhomogeneous nanostructures, and also to prepare others having very broadly multimodal chain‐length distributions. Macroscopic phase separation, probably high crosslink density clusters, was observed to occur in some of the bimodal networks. The mechanical properties in simple extension and in equilibrium swelling were measured. The bimodal elastomers that were not obviously inhomogeneous showed very good mechanical properties, but the macroscopically phase‐separated networks, and the broadly multimodal network were weak. Analysis of the Mooney‐Rivlin profiles suggests that the reinforcing mechanism could have a structural component in addition to that from the limited extensibilities of the short chains. The mechanical properties and the extents of swelling support the cluster conjecture, in accord with previous morphological studies on spatially‐inhomogeneous polysiloxane elastomers.     

Valence state parameters of all transition metal atoms in metalloproteins—development of ABEEMσπ fluctuating charge force field

To promote accuracy of the atom‐bond electronegativity equalization method (ABEEMσπ) fluctuating charge polarizable force fields, and extend it to include all transition metal atoms, a new parameter, the reference charge is set up in the expression of the total energy potential function. We select over 700 model molecules most of which model metalloprotein molecules that come from Protein Data Bank. We set reference charges for different apparent valence states of transition metals and calibrate the parameters of reference charges, valence state electronegativities, and valence state hardnesses for ABEEMσπ through linear regression and least square method. These parameters can be used to calculate charge distributions of metalloproteins containing transition metal atoms (Sc‐Zn, Y‐Cd, and Lu‐Hg). Compared the results of ABEEMσπ charge distributions with those obtained by ab initio method, the quite good linear correlations of the two kinds of charge distributions are shown. The reason why the STO‐3G basis set in Mulliken population analysis for the parameter calibration is specially explained in detail. Furthermore, ABEEMσπ method can also quickly and quite accurately calculate dipole moments of molecules. Molecular dynamics optimizations of five metalloproteins as the examples show that their structures obtained by ABEEMσπ fluctuating charge polarizable force field are very close to the structures optimized by the ab initio MP2/6–311G method. This means that the ABEEMσπ/MM can now be applied to molecular dynamics simulations of systems that contain metalloproteins with good accuracy. © 2014 Wiley Periodicals, Inc.     

基于机器学习的直接电离质谱爆炸物检测方法

直接电离质谱系统在现场快速检测中的应用日益广泛,主要用于爆炸物、毒品、食品添加剂等的检测。然而,直接电离质谱系统中质谱信号波动大且同一浓度样品峰强呈现对数正态分布,严重影响了检出限附近低浓度样品的检测准确性。该研究将乙酰水杨酸(115个样品)作为爆炸物模拟物,利用介质阻挡放电离子源与质谱系统,研究了基于机器学习的直接电离质谱数据预处理和分类算法,以提高低浓度样品的检测准确率。对两种浓度为1 ng/mL的常见爆炸物样本(三硝基甲苯和硝酸铵分别为110、90个)及空白对照样本(366个)开展了应用实验。结果表明,与传统提取离子流方法和高斯混合模型方法相比,采用随机森林算法可将F_score从0.74、0.89提升至0.96,显著提高了检测准确率,且单个样本数据分析时间远少于0.1 s,满足实时检测需求。     

Automatic analysis of microscopic photo-pictures of undeveloped and developed nuclear emulsions

Methods of automatic computer analyses of images have large impotence in numerous physical problems with different kinds of radiations, which use such solid-state detectors as nuclear photographic emulsions, plastics and others. We described a method of expanding boundaries of dark areas for such computer analyses of micro-pictures. We demonstrated effectiveness of this method by the example of the search of microcrystals size distributions in two undeveloped nuclear emulsions and similar distribution of background grains in the developed emulsion.     

Direct Derivation of Free Energies of Membrane Deformation and Other Solvent Density Variations From Enhanced Sampling Molecular Dynamics

We report a methodology to calculate the free energy of a shape transformation in a lipid membrane directly from a molecular dynamics simulation. The bilayer need not be homogeneous or symmetric and can be atomically detailed or coarse grained. The method is based on a collective variable that quantifies the similarity between the membrane and a set of predefined density distributions. Enhanced sampling of this “Multi-Map” variable re-shapes the bilayer and permits the derivation of the corresponding potential of mean force. Calculated energies thus reflect the dynamic interplay of atoms and molecules, rather than postulated effects. Evaluation of deformations of different shape, amplitude, and range demonstrates that the macroscopic bending modulus assumed by the Helfrich–Canham model is increasingly unsuitable below the 100-Å scale. In this range of major biological significance, direct free-energy calculations reveal a much greater plasticity. We also quantify the stiffening effect of cholesterol on bilayers of different composition and compare with experiments. Lastly, we illustrate how this approach facilitates analysis of other solvent reorganization processes, such as hydrophobic hydration. Published 2019. This article is a U.S. Government work and is in the public domain in the USA.     

Parametric Duality Models for Semiinfinite Multiobjective Fractional Programming Problems Containing Generalized （α,η,ρ）-V-Invex Functions

In this paper, we present several parametric duality results under various generalized ( α,η,ρ )-V- invexity assumptions for a semiinfinite multiobjective fractional programming problem.     

A probabilistic analysis of a discrete-time evolution in recombination

We study the discrete-time evolution of a recombination transformation in population genetics. The transformation acts on a product probability space, and its evolution can be described by a Markov chain on a set of partitions that converges to the finest partition. We describe the geometric decay rate to this limit and the quasi-stationary behavior of the Markov chain when conditioned on the event that the chain does not hit the limit.