A framework for classification with single feature kernel matrix

In this paper, we propose a novel classification framework using single feature kernel matrix. Different from the traditional kernel matrices which make use of the whole features of samples to build the kernel matrix, this research uses features of the same dimension of any two samples to build a sub-kernel matrix and sums up all the sub-kernel matrices to get the single feature kernel matrix. We also use single feature kernel matrix to build a new SVM classifier, and adapt SMO (Sequential Minimal Optimization) algorithm to solve the problem of SVM classifier. The results of the experiments on several artificial datasets and some challenging public cancer datasets display the classification performance of the algorithm. The comparisons between our algorithm and L₂-norm SVM on the cancer datasets demonstrate that the accuracy of our algorithm is higher, and the number of support vectors selected is fewer, indicating that our proposed framework is a more practical approach.     

近红外光谱的北方寒地土壤含水率预测模型研究

我国北方寒地温差大,土壤温差对近红外光谱测量土壤墒情有较大影响。针对这一问题,以北方寒地土壤为研究对象,探究大范围温度胁迫下（-20～40 ℃）土壤的近红外光谱与土壤不同含水率之间的关系预测模型方法。选取黑龙江八一农垦大学农学院试验基地中的黑土,经烘干、过筛等操作处理后配置含水率范围在15%～50%内八种不同湿度的土壤样品,建立北方寒地土壤大范围温度胁迫下土壤的近红外光谱信息与含水率之间的定量预测模型。在全波段光谱数据的基础上,结合五种不同光谱信号预处理方法,采用BP神经网络算法、优化支持向量机算法（SVM）、高斯过程算法（GP）三种智能算法建立北方寒地土壤近红外光谱与含水率的预测模型并验证模型的效果。利用69组数据进行训练建模, BP神经网络相关参数设置为学习速率0.05,最大训练次数设置为5 000,隐层单元数确定为20;SVM采用径向基函数,并利用leave-one-out cross validation确定了最佳惩罚参数为0.87,使模型预测的准确性提高;高斯过程算法内部采用马顿核。模型的定量评估采用决定系数(R2)和均方根误差（RMSE）。结果表明,在建立的全部BP神经网络模型中,效果最佳的为S_G-BP神经网络模型,模型的R2为0.960 9,RMSE为2.379 7;在SVM模型中SNV-SVM模型的效果最好,模型的R2为0.991 1,RMSE为1.081 5;在GP模型中S_G-GP模型的效果最好,模型的R2为0.928,RMSE为3.258 1,综上基于SNV预处理的SVM模型训练效果最优。利用剩余的35组光谱数据作为预测集验证模型性能,经模型对比分析发现基于SVM算法的预测模型效果优于其他两种算法,其中基于S_G的SVM模型效果最优,其预测模型的R2和差RMSE分别为0.992 1和0.736 9。综合建模集与预测集的参数最终确定基于S_G的SVM模型为最佳模型。此模型可以作为大范围温度胁迫条件下（寒地）的土壤含水率有效预测方法,为设计优化适宜寒地便携式近红外土壤含水率快速测量仪提供科学依据。     

Nonlinearity effects in multidimensional scaling

When multidimensional scaling of n cases is derived from dissimilarities that are functions of p basic continuous variables, the question arises of how to relate the values of the variables to the configuration of n points. We provide a methodology based on nonlinear biplots that expresses nonlinearity in two ways: (i) each variable is represented by a nonlinear trajectory and (ii) each trajectory is calibrated by an irregular scale. Methods for computing, calibrating and interpreting these trajectories are given and exemplified. Not only are the tools of immediate practical utility but the methodology established assists in a critical appraisal of the consequences of using nonlinear measures in a variety of multidimensional scaling methods.     

基于平稳小波和相空间重构的激光混沌预测

提出了一种激光混沌时间序列预测算法.该算法通过平稳小波分解，将原始数据序列分解为与原序列等长的尺度系数和小波系数，利用坐标延迟理论，重建各级尺度系数和各级小波系数的相空间,再根据混沌吸引子的稳定性和分形性,在相空间中对尺度系数和小波系数进行预测,进而通过平稳小波重构算法,实现了时间序列的非线性预测.该算法对数据可以进行更平滑的处理，比无小波算法预测的时间范围更长.通过仿真试验说明，原始时间数据序列被成功的重建，说明算法能够有效的对非线性动态系统的时间序列进行建模和预测.     

不同加载方式下路堤沉降性状与预测模型研究

从考虑土体流变的本构关系出发 ,证明了瞬时加载、线性或者近似线性加载和分级加载情况下路堤沉降的性状 ,提出了从数学上能满足各自变形特征的预测模型 ,工程实例分析表明了在不同条件下采用不同模型的合理性和实用性     

Azeotropic behaviour of (benzene + cyclohexane + chlorobenzene) ternary mixture using chlorobenzene as entrainer at 101.3 kPa

In this paper, the azeotropic behaviour of the (benzene + cyclohexane + chlorobenzene) ternary mixture was experimentally investigated with the aim of enhancing the knowledge for the feasible use of chlorobenzene as an entrainer for the azeotropic distillation of the binary azeotrope. Such a study has not been reported in the literature to the best of the authors’ knowledge. (Vapour + liquid) equilibria data for (benzene + cyclohexane + chlorobenzene) at 101.3 kPa were obtained with a Othmer-type ebulliometer. Data were tested and considered thermodynamically consistent. The experimental results showed that this ternary mixture is completely miscible and exhibits an unique binary homogeneous azeotrope, an unstable node at the conditions studied, and the propitious topological characteristics (residual curve map and relative volatility) to be separated. Satisfactory results were obtained for the correlation of equilibrium compositions with the UNIQUAC activity coefficients model and also for prediction with the UNIFAC method. In both cases, low root mean square deviations of the vapour mole fraction and temperature were calculated. The capability of chlorobenzene as a modified distillation agent at atmospheric condition is discussed in terms of the thermodynamic topological analysis. A conceptual distillation scheme with reversed volatility is proposed to separate the azeotropic mixture. In order to reduce the operational cost requirements of the sequence of columns proposed, the range for optimal reflux and the ratio for feed flow conditions were studied.     

Prediction of reverberation time using the residual minimization method

In many practical situations the assumption of sound field dispersion needed for the application of the Sabine’s theory is not fulfilled. In general, sound field is sufficiently dispersed if there are no large differences in the dimensions of the room, limiting partitions are not parallel, or the sound absorbing material is uniformly distributed. In practice, very few of these requirements are satisfied. As a result, a number of other formulas describing reverberation time have been created, for example Fitzroy’s or Neubauer’s formulas. However, these methods in many cases differ significantly from the actual measurements. The paper presents a method used to estimate reverberation time as well as its applicability potential involving laboratory models and auditorium rooms. The proposed method can be classified into a group of learning methods and involves the use of statistical methods which allow for approximation with the use of the least squares method.     

用遗传算法拟合电力系统负荷短期预测的非线性模型

本文运用遗传算法，根据河北省１９８５—１９９０年春节期间每小时用电量的统计数字，建立了非线性回归模型（Ｇｏｍｐｅｒｔｄｚ），结果表明，遗传算法性能良好，可望成为电力系统各种非线性模型辨识的有效手段。     

Prediction of vapor-liquid equilibrium from excess enthalpy data for binary ether +n-alkane or cyclohexane mixtures

Vapor liquid equilibrium (VLE) is successfully predicted from excess enthalpy H^E data for binary ether + n-alkane or cyclohexane mixtures. Parameters for the continuous linear association model (CLAM) and for the UNIQUAC Model for the excess Gibbs energy G^E were determined from H^E data measured at a low temperature (ambient temperature). These parameters are used to predict VLE data at low and high temperatures. The dependence of the accuracy of predictions on the set of H^E data chosen to evaluate the parameters and on the model for G^E are discussed.     

关于预测误差平方和最小准则的几点看法

在应用桃棣荣等提出的预测误差平方和（ＰＲＥＳＳ）最小准则逐步算法的过程中，发现单纯应用ＰＲＥＳＳ值来控制引进因子数目有时不很有效。因此建议：１）当引进因子过多时，采用赤池信息准则对因子数做进一步控制。２）当引进因子过少时，取消原算法中关于ＰＲＥＳＳ值上界的规定和增加候选因子数以增加引进因子数。最后，对应用ＰＲＥＳＳ准则时所选因子的优良性进行了讨论