基于Kronecker型替换的黑盒多项式稀疏插值

本文基于新的Kronecker型替换,给出两个由黑盒表示的稀疏多项式的新确定性插值算法.令f∈R[x1,……,xn]是一个稀疏黑盒多项式,其次数上界为D.当R是C或者是有限域时,相对于已有算法,新算法具有更好的计算复杂度或者关于D的复杂度更低.特别地,对于一般黑盒模型,D是复杂度中的主要因素,而在所有的确定性算法中,本文的第二个算法的复杂度关于D是最低的.     

Qualitative and quantitative determination of Atractylodes rhizome using ultra‐performance liquid chromatography coupled with linear ion trap–Orbitrap mass spectrometry with data‐dependent processing

A quick and effective workflow based on ultra‐performance liquid chromatography coupled with electron spray ionization and LTQ‐Orbitrap mass spectrometry (UPLC‐LTQ‐Orbitrap MS) was established for compositional analysis and screening of the characteristic compounds of three species of Atractylodes rhizome for quality evaluation. This technique was employed to determine the seven main components in Atractylodes rhizome samples. Ultimately, 78 constituents were identified; of these, seven characteristic compounds were selected for species discrimination, comprising atractylodin (63), atractylenolide I (43), atractylenolide II (49), atractylenolide III (53), atractylon (69), methyl‐atractylenolide II (54) and (4E,6E,12E)‐tetradecadecatriene‐8,10‐diyne‐1,3‐diacetate (59). The seven main compounds, including six characteristic compounds, were simultaneously determined in 29 batches of Atractylodes rhizome samples. Thus, the method validation showed acceptable results. Quantitative analysis showed significantly different contents of the seven main components among the three species of Atractylodes rhizome, which indicates possible distinctions in the pharmacological effects. This established method can simultaneously provide qualitative and quantitative results for compositional characterization of Atractylodes rhizomes and for quality control.     

Preparation and characterization of new inorganic–organic hybrid catalyst H3PMo12O40/Hyd‐SBA‐15 and its application in the domino multi‐component reaction

We present novel inorganic–organic hybrid catalyst to accomplish domino multi‐component reaction (MCR) for synthesis of 3‐amino‐2′‐oxospiro[benzo[c]pyrano[3,2‐a]phenazine‐1,3′‐indoline]‐2‐carbonitrile/carboxylate derivatives. This methodology offers remarkable development by easy production of H₃PMo₁₂O₄₀/Hyd‐SBA‐15 in regard to solving the problem of using harsh catalysts, also it demonstrates to be impressive and environmentally friendly in term of low reaction times and high yields.     

Sparse Multicategory Generalized Distance Weighted Discrimination in Ultra-High Dimensions

Distance weighted discrimination (DWD) is an appealing classification method that is capable of overcoming data piling problems in high-dimensional settings. Especially when various sparsity structures are assumed in these settings, variable selection in multicategory classification poses great challenges. In this paper, we propose a multicategory generalized DWD (MgDWD) method that maintains intrinsic variable group structures during selection using a sparse group lasso penalty. Theoretically, we derive minimizer uniqueness for the penalized MgDWD loss function and consistency properties for the proposed classifier. We further develop an efficient algorithm based on the proximal operator to solve the optimization problem. The performance of MgDWD is evaluated using finite sample simulations and miRNA data from an HIV study.     

Normal halogen dependence of 13C NMR chemical shifts of halogenomethanes revisited at the four‐component relativistic level

The ‘Normal Halogen Dependence’ of ¹³C NMR chemical shifts in the series of halogenomethanes is revisited at the four‐component relativistic level. Calculations of ¹³C NMR chemical shifts of 70 halogenomethanes have been carried out at the density functional theory (DFT) and MP2 levels with taking into account relativistic effects using the four‐component relativistic theory of Dirac‐Coulomb within the different computational methods (4RPA, 4OPW91) and hybrid computational schemes (MP2 + 4RPA, MP2 + 4OPW91). The most efficient computational protocols are derived for practical purposes. Relativistic shielding effect reaches as much as several hundreds of ppm for heavy halogenomethanes, and to account for this effect in comparison with experiment at the qualitative level, relativistic Dyall's basis sets of triple‐zeta quality or higher are to be used within the framework of the four‐component relativistic theory taking into account solvent effects. Relativistic geometrical optimization (as compared with the non‐relativistic level) is essential for the molecules containing at least two iodines at one carbon atom. Copyright © 2016 John Wiley & Sons, Ltd.     

Multiple bright and dark soliton solutions in three component spinor Bose-Einstein condensates

We investigate a quasi one dimensional spin-1 Bose-Einstein Condensates (BEC) in the absence of an external confinement governed by a system of three coupled Gross-Pitaevskii (GP) equation. Based on the Lax-pair, we construct one soliton solution employing gauge transformation method. In addition, the multiple bright and dark soliton solutions are obtained by properly choosing amplitude dependent parameter in the Lax-pair. The results of the paper emphasizes the richness in the structure of soliton solutions admitted by the spin components, a phenomenon which has never been brought out to the fore. We have also extended the gauge transformation method to generate two soliton solutions.     

Metabolites identification of oil palm roots infected with Ganoderma boninense using GC–MS-based metabolomics

An approach to metabolomics profiling of non-infected and Ganoderma boninense (G. boninsense) infected oil palm roots crude extracts that utilize gas chromatography-mass spectrometry (GC–MS) and multivariate statistics of principal component analysis (PCA) have been tested. This combination has provided a rapid approach in investigating the changes in the metabolite variations of non-infected and infected oil palm roots at 14 and 30 days post-infection. The extracts were prepared by using 80% (v/v) of methanol. In identifying the metabolites responsible for each differentiation, PCA model was generated in loading bi-plot. Dimethyl benzene-1,4-dicarboxylate, methyl 3-(3,5-ditert-butyl-4-hydroxyphenyl)propanoate, ergost-5-en-3-ol, (3β), stigmast-5-en-3-ol, (3β), stigmasterol, methyl hexadecanoate, methyl (9Z,12Z)-octadeca-9,12-dienoate, methyl octadecanoate, 2-(hydroxymethyl)-2-nitropropane-1,3-diol, methyl (Z)-octadec-6-enoate and (E)-icos-5-ene were found more abundant in G. boninense infected roots than in non-infected roots. Steroidal compounds and fatty acid derivatives which has been determined in the non-infected and G. boninense infected roots regulate a variety of responses to the G. boninense. The abundant of these metabolites in G. boninense infected roots are due to the crucial roles in pathogen defence.     

卷积神经网络用于近红外光谱古筝面板木材分级

目前,我国乐器制作行业在古筝面板用木材等级的筛选上主要依赖于技师主观评判,但此法缺少科学理论的依据,效率低,客观性及出材率的提高等方面受到限制,无法满足乐器市场的大量需求。实现古筝面板用木材快速、智能化的分级工作是一个急需解决的课题。近红外光谱非常适用于测量含氢的有机物质。古筝面板木材主要化学成分的化学键均由含氢基团组成,不同等级板材的化学成分存在差异,这些差异反映在近红外光谱中,为判断木材等级提供了可能。同时卷积神经网络对非线性数据具有较强的特征提取能力,所以提出一种应用卷积神经网络模型对光谱数据进行分析的方法,进而判别木材的等级。应用了Savitzky Golay一阶、二阶微分两种预处理方法和核主成分分析、连续投影算法两种数据压缩方法,通过所设计的卷积神经网络模型以样本识别准确率和模型构建过程中的损失值作为判定指标选出最佳预处理和数据压缩方法。为了提高模型提取分析光谱数据的能力和避免过拟合现象,应用了多通道卷积核、批量归一化和early stopping策略,将通过两层卷积层提取的特征信息送入全连接层,从而充分提取剩余信息,通过Softmax函数获得板材的最终预测等级,从而确定了最终模型。最终Savitzky Golay一阶微分和核主成分分析为最佳数据处理方法,同时得出用于区分不同等级的古筝面板用木材的主要关键谱带,分别为1 163~1 243, 1 346~1 375和1 525~1 584 nm。将该模型应用于测试集样本,古筝面板用木材的等级识别准确率为95.5%。实验结果表明所提出的方法可以高效地处理光谱数据,有效识别区分不同等级的古筝面板用木材的关键特征,从而为广阔的乐器市场提供一定的技术支持。