人工神经网络-荧光光谱法同时测定维生素B1,B2,B6

应用人工神经网络原理 ,以Levenberg MarquardtBP算法对荧光光谱严重重叠的维生素B1 ,B2 ,B6三组份混合体系同时进行了含量测定。在 390～ 5 2 0nm的范围内 ,以 15个特征波长处的荧光强度值作为网络特征参数 ,并通过均匀设计安排样本 ,经网络训练和计算得出维生素B1 ,B2 ,B6 三者的平均回收率分别为 99 86 % ,99 6 0 % ,99 4 9% ,测定结果的相对标准偏差各为 1 7% ,1 6 % ,1 7%。     

Is the predictability of emerging and developed stock markets really exploitable?

A number of recent papers have analyzed the degree of predictability of stock markets. In this paper, we firstly study whether this predictability is really exploitable and secondly, if the economic significance of predictability is higher or lower in the emerging stock markets than in the developed ones. We use a variety of linear and nonlinear – Artificial Neural Networks – models and perform a computationally demanding forecasting experiment to assess the predictability of returns. Since we are interested in comparing the predictability in economic terms we also propose a modification in the nets’ loss function for market trading purposes. In addition, we consider both explicit and implicit trading costs for emerging and developed stock markets. Our conclusions suggest that, in contrast to some previous studies, if we consider total trading costs both the emerging as well as the developed stock returns are clearly nonpredictable. Finally, we find that Artificial Neural Networks do not provide superior performance than the linear models.     

The numerical solution of linear ordinary differential equations by feedforward neural networks

It is demonstrated, through theory and examples, how it is possible to construct directly and noniteratively a feedforward neural network to approximate arbitrary linear ordinary differential equations. The method, using the hard limit transfer function, is linear in storage and processing time, and the L₂ norm of the network approximation error decreases quadratically with the increasing number of hidden layer neurons. The construction requires imposing certain constraints on the values of the input, bias, and output weights, and the attribution of certain roles to each of these parameters.

All results presented used the hard limit transfer function. However, the noniterative approach should also be applicable to the use of hyperbolic tangents, sigmoids, and radial basis functions.     

Site specific biotinylation of the human aldo/keto reductase AKR1A1 for immobilization

New strategies for the specific monolabelling of enzymes play a key role in the development of artificial proteins. Especially for the emerging research field of nanobiotechnology and bioelectronics artificial monofunctionalized redoxproteins are of great interest. The human AKR1A1, an enzyme of the aldo/keto reductase superfamily, has been chosen as subject for the synthesis of an artificially mono biotinylated redoxprotein in order to selectively immobilize this enzyme for bioelectronic applications. To produce monofunctionalized enzyme we applied the strategy of Expressed Protein Ligation (EPL) in combination with solid phase peptide synthesis (SPPS). Accordingly, we used the IMPACT^®-system and cloned the aldo/keto-reductase as fusion protein with an additional intein/chitin binding domain. Through intein mediated splicing we could produce the C-terminal thioester of the aldo/keto-reductase, which maintained its biological activity. Then, the thioester was coupled to Cys-Lys(Ahx-Ahx-biotin)-amide by Native Chemical Ligation, which led to mono-biotinylated protein. The enzyme activity was proven to be intact as shown by various kinetic investigations. Immobilization was performed on avidin coated silica microspheres. Accordingly, for the first time selectively modified AKR1A1 has been immobilized.     

人工神经网络在建筑声学中的应用

本文对人工神经网络的历史作了简单回顾，并对它在建筑声学领域中的应用状况作了阐述，同时提出了有待进一步探讨的若干问题。     

Effective downstream boundary conditions for incompressible Navier–Stokes equations

The aim of this paper is to give open boundary conditions for the incompressible Navier–Stokes equations. From a weak formulation in velocity–pressure variables, some natural boundary conditions involving the traction or pseudotraction and inertial terms are established. Numerical experiments on the flow behind a cylinder show the efficiency of these conditions, which convey properly the vortices downstream. Comparisons with other boundary conditions for the velocity and pressure are also performed.     

Structure Elucidation and Artificial Intelligence

The difference between the older and the more recent computer systems amounts, in the final analysis, to the difference between an insect and a human being. The insect comes “fully programmed” into the world–programmed by instincts which it follows without using any thought processes. On the other hand, the human being must first learn the correct behavior–however, such learning can develop an impetus all its own: becoming, indeed, proceeding from his own decision and his awareness, the human being can change his former action programs. The creation of a computer system of the X -th generation will actually far more resemble educating a child than the traditional programming of a computer.     

Hopfield神经网络中的数学问题

Ｈｏｐｆｉｅｌｄ人工神经网络和多层网络模型中有许多基本的数学问题，其中，最重要的是存储容量的问题，即吸引子的个数，在以往的文章中，大多数都以概率模型为背景，本文从组合论的角度提供了几个问题，我们把Ｈｏｐｆｉｅｌｄ网络中的状态按连接矩阵所构成的变换分类，网络的容量就是类的个数，我们得出了一些基本结果，还有一些有意义的问题，如多层网络中的结果是否可交换等问题，还有待进一步研究。     

空间人造目标混合光谱材料数目确定方法仿真研究

在空间人造目标光谱分析领域,受到观测距离和观测设备空间分辨率的限制,通常在观测空间人造目标光谱信号时,目标某个瞬时视场中的多种纯物质材料的光谱特征信息组合在一个像元中,形成“混合光谱”。因此,将这些混合光谱分解为每个单一材料的光谱并估计出相应的组成比例是空间人造目标光谱分析研究的重点。大多数现有空间目标光谱分解方法都假设空间人造目标混合光谱中包含的纯物质材料种类个数（即“端元数目”）是先验已知的,这对于未知空间人造目标而言是不现实的。因此,纯物质材料数目正确估计对后续光谱数据分析处理的准确性起着至关重要的作用。目前,现有的端元数目确定方法的设计均在高斯白噪声的假设下进行,而对于噪声信号的分布存在频谱相关性的情况下,会提供较差的结果。采用一种基于数据内在维度和似然最大化理论的方法--鲁棒特征值极大似然方法。由于数据内在维数与信号协方差矩阵和信号相关矩阵特征值差异的统计分布特性高度相关,因此通过分析该特征值差异的统计分布特性,构建一个极大似然函数,可以实现空间人造目标混合光谱端元数目的确定。该方法包含两个步骤：首先,采用基于多元回归和改进最小噪声分离方法对原始光谱数据进行预处理完成噪声特性估计和噪声白化过程,从而有效抑制具有频谱相关性的噪声的干扰;接下来,通过求解一个离散对数联合似然函数的极大值问题来实现空间人造目标混合光谱端元数目的确定,该方法完全不需要输入任何参数,并且运行速度比较快。分别利用实验室实测的五种空间人造目标材料的可见/近红外光谱数据和美国地质勘测局光谱数据构建混合光谱仿真数据进行实验。结果表明,该方法能有效抑制相关噪声和白噪声的干扰,空间人造目标纯物质材料数目确定结果具有很好的准确性和稳定性。     

牛肉化学成分的近红外光谱检测方法的研究

通过对整块牛肉和肉馅样品进行扫谱,测定其脂肪、蛋白和水分含量,采用国产 SupNIR-1000 近红外光谱仪,运用人工神经网络(ANN)分别建立肉馅和整块牛肉的脂肪、蛋白和水分的模型.肉馅样品的脂肪模型校正相关系数为0.971、预测相关系数为0.972;蛋白的校正相关系数为0.952、预测相关系数为0.949;水分的校...