Ab initio Tertiary Structure Prediction of Proteins

A daunting challenge in the area of computational biology has been to develop a method to theoretically predict the correct three-dimensional structure of a protein given its linear amino acid sequence. The ability to surmount this challenge, which is known as the protein folding problem, has tremendous implications. We introduce a novel ab initio approach for the protein folding problem. The accurate prediction of the three-dimensional structure of a protein relies on both the mathematical model used to mimic the protein system and the technique used to identify the correct structure. The models employed are based solely on first principles, as opposed to the myriad of techniques relying on information from statistical databases. The framework integrates our recently proposed methods for the prediction of secondary structural features including helices and strands, as well as -sheet and disulfide bridge formation. The final stage of the approach, which culminates in the tertiary structure prediction of a protein, utilizes search techniques grounded on the foundations of deterministic global optimization, powerful methods which can potentially guarantee the correct identification of a protein's structure. The performance of the approach is illustrated with bovine pancreatic trypsin inhibitor protein and the immunoglobulin binding domain of protein G.     

Approximation and Prediction of the Numerical Solution of Some Burgers Problems

The Aitken's ²-prediction of Brezinski has already been used by Morandi Cecchi et al. in order to compute a numerical approximation of the solution of a parabolic initial-boundary value problem. This method consists in two consecutive steps: the first one is the approximation with a finite elements method, where the solution of the involved nonlinear system is computed by Gauss–Seidel method; the second one is a prediction of further terms with Aitken's ²-process. By comparison with this method, we use other methods of prediction in another way. First, we consider a generalization of ²-prediction, the so-called -prediction. In this paper, we only use vector prediction which is more stable than the scalar one. Then, the methods of prediction presented can be used in order to predict the starting vector of the Gauss–Seidel method.     

Refinement of modelled structures by knowledge-based energy profiles and secondary structure prediction: Application to the human procarboxypeptidase A2

Knowledge-based energy profiles combined with secondary structure prediction have been applied to molecular modelling refinement. To check the procedure, three different models of human procarboxypeptidase A2 (hPCPA2) have been built using the 3D structures of procarboxypeptidase A1 (pPCPA1) and bovine procarboxypeptidase A (bPCPA) as templates. The results of the refinement can be tested against the X-ray structure of hPCPA2 which has been recently determined. Regions miss-modelled in the activation segment of hPCPA2 were detected by means of pseudo-energies using Prosa II and modified afterwards according to the secondary structure prediction. Moreover, models obtained by automated methods as COMPOSER, MODELLER and distance restraints have also been compared, where it was found possible to find out the best model by means of pseudo-energies. Two general conclusions can be elicited from this work: (1) on a given set of putative models it is possible to distinguish among them the one closest to the crystallographic structure, and (2) within a given structure it is possible to find by means of pseudo-energies those regions that have been defectively modelled.     

Bayesian networks applied to credit scoring

Email: kchang{at}gmu.eduEmail: RobertFung{at}Fairlsaac.comEmail: alan.lucas{at}hotmail.com^¶Email: BobOliver{at}Fairlsaac.com^||Email: NShikaloff{at}Fairlsaac.com The objectives of this paper are to apply the theory and numericalalgorithms of Bayesian networks to risk scoring, and comparethe results with traditional methods for computing scores andposterior predictions of performance variables. Model identification,inference, and prediction of random variables using Bayesiannetworks have been successfully applied in a number of areas,including medical diagnosis, equipment failure, informationretrieval, rare-event prediction, and pattern recognition. Theability to graphically represent conditional dependencies andindependencies among random variables may also be useful incredit scoring. Although several papers have already appearedin the literature which use graphical models for model identification,as far as we know there have been no explicit experimental resultsthat compare a traditionally computed risk score with predictionsbased on Bayesian learning algorithms. In this paper, we examine a database of credit-card applicantsand attempt to ‘learn’ the graphical structure ofthe characteristics or variables that make up the database.We identify representative Bayesian networks in a developmentsample as well as the associated Markov blankets and cliquestructures within the Markov blanket. Once we obtain the structureof the underlying conditional independencies, we are able toestimate the probabilities of each node conditional on its directpredecessor node(s). We then calculate the posterior probabilitiesand scores of a performance variable for the development sample.Finally, we calculate the receiver operating characteristic(ROC) curves and relative profitability of scorecards basedon these identifications. The results of the different modelsand methods are compared with both development and validationsamples. Finally, we report on a statistical entropy calculationthat measures the degree to which cliques identified in theBayesian network are independent of one another.     

基于搜索平均法的气象观测数据的非线性去噪

介绍了一种新型基于相空间重构和返回图法的非线性降噪方法——搜索平均法，首先将该方法应用于被高斯白噪声所污染的Henon映射时间序列，说明根据降噪理论所编写的计算程序的正确性；然后应用于中国气象局公布全国435站1960—2000年的逐日气温观测时间序列，并利用非线性预报方法衡量降噪的效果，验证了该方法的有效性.     

基于时变阈值过程神经网络的太阳黑子数预测

太阳黑子活动直接影响着外层空间环境的变化，为保证航天飞行任务的安全必须对其进行有效预测.为此，提出了一种基于时变阈值过程神经网络的时间序列预测模型.为简化模型的计算复杂度，开发了一种基于正交基函数展开的学习算法.文中分析了模型的泛函逼近能力，并以Mackey-Glass时间序列预测为例验证了所提模型及其学习算法的有效性.最后，将该预测模型用于太阳活动第23周太阳黑子数平滑月均值预测，取得了满意的结果，应用结果同时表明：所提预测方法与其他传统预测方法相比预测精度有所提高，具有一定的理论和实用价值. 关键词： 太阳黑子数 时变阈值过程神经网络 时间序列预测 泛函逼近     

线型优化最大熵线性预测方法自回归模型三种求解方法的比较

采用三种方法：修正协方差法MCOV(Modified Covariance Method)、递推极大似然估计法RMLE(Recursive Maximum Likelihood Estimator)和伯格法(Burg Method)求解线型优化最大熵线性预测方法中的自回归模型系数,并且在不同求解方法情况下,将阶次、信噪比对光谱复原的影响作了详尽的比较.研究结果表明:在线型优化最大熵线性预测方法的自回归模型系数三种求解方法中,修正协方差法要优于递推极大似然估计法和伯格法.     

Discussion of Some Problems About Nonlinear Time Series Prediction Using v-Support Vector Machine

Some problems in using v-support vector machine (v-SVM) for the prediction of nonlinear time series are discussed. The problems include selection of various net parameters, which affect the performance of prediction, mixture of kernels, and decomposition cooperation linear programming v-SVM regression, which result in improvements of the algorithm. Computer simulations in the prediction of nonlinear time series produced by Mackey-Glass equation and Lorenz equation provide some improved results.     

半色调荧光图像的光谱反射与透射模型

荧光油墨半色调印刷品的显色预测规律是彩色成像领域内十分关键的课题。把荧光半色调印刷品反射出来的光分成两个独立的部分,即由最初入射光组成的主光流和由于吸收了主光流而产生的荧光流;并且采用了一个指数矩阵来描述墨层产生的荧光光强,从而得出了荧光半色调图像色彩的光反射规律。考虑到入射光中从一种颜色油墨入射后再由同色油墨出射的概率要比从其它色油墨出射的概率要大的事实,讨论中引入权重因子描述从同色油墨出射部分与从整个表面出射部分比例不同的现象,建立了荧光油墨半色调图像的Clapper-Yule新光谱反射率模型。     

开式轴流风扇气动噪声预测

本文采用LES/FW-H的匹配方法,研究了开式轴流风扇内部旋涡流动特征及其与叶片表面干涉引起的气动噪声之间的联系,同时进行了远场噪声预测,探讨了叶轮不同表面辐射噪声时的频谱分布特征.研究结果表明,开式轴流风扇吸力面附近形成的叶尖涡和前缘分离涡在吸力面叶片表面相应位置形成大压力波动,形成主要噪声源;叶片吸力面的辐射噪声可以通过改善吸力面附近的旋涡流动来降低;低速轴流叶轮由叶轮壁面辐射的噪声以宽频成分为主.