基于全局数据混洗的小样本数据预测方法

以广州车牌竞拍价格数据集为数据来源, 采用线性回归并结合k折交叉验证, 研究小样本数据集的预测方法。为解决小样本局部特异性数据导致的验证误差增大的问题, 提出验证之前先对数据进行全局混洗的策略。最后通过实验验证了此策略可以明显降低验证误差, 以此为基础, 通过多组实验验证, 确定了合适的参数, 结果表明最终预测值的总平均正确率达到了95%。     

Application Research of Robust LS-SVM Regression Model in Forecasting Patent Application Counts

A forecasting system of patent application counts is studied in this paper. The optimization model proposed in the research is based on support vector machines (SVM), in which cross-validation algorithm is used for preferences selection. R esults of data simulation show that the proposed method has higher forecasting p recision power and stronger generalization abi1ity than BP neural network and RB F neural network. In addition, it is feasible and effective in forecasting paten t application counts.     

边坡稳定性预测的Bayes判别分析方法及应用

边坡稳定性的分析是一个复杂的系统工程问题.基于Bayes判别分析(BDA)理论并结合工程实际,选用边坡岩体的重度黏聚力、摩擦角、边坡角、边坡高度及孔隙压力比等6个指标作为边坡稳定性预测的判别因子,建立边坡稳定性预测的Bayes判别分析模型;以32组边坡实测数据作为学习样本进行训练,建立Bayes线性判别函数;以交差确认估计法对判别准则进行评价以检验模型的优良性,以Bayes线性判别函数计算7个待判样品的Bayes判别函数值.研究表明:Bayes判别分类性能良好,与支持向量机方法有较好的一致性,且预测精度高,交差确认估计的误判率较低,为边坡稳定性预测提供了一种新思路.     

红外与可见光复合寻的制导中的快速图像配准方法

对现有多传感器图像配准方法在红外与可见光复合寻的制导应用中存在的计算量大等不足,提出了一种基于部分Hausdorff距离的快速方法和一种两集合相似度的定义,从而实现了有效减小误匹配特征点对的红外与可见光图像快速配准方法.此算法通过分别提取两幅图像边缘上显著特征点,再根据提出的快速方法和两集合相似度定义,粗略估计出图像间的旋转角度、平移量;然后依据此估计值,缩小特征点匹配时匹配特征点搜索范围,求得一致特征点对;最后用最小二乘法求解最优变换参数.实验结果表明了该方法的有效性、快速性.     

精馏塔系的软测量建模与优化策略研究

针对工业精馏系统操作优化的需求,在分析过程机理的基础上,本文基于广义回归神经网络和交叉验证等方法,建立了产品质量的软测量模型。随后,通过改进标准遗传算法的算子和结构,提出一种改进的求解有约束稳态优化问题的遗传算法,解决了常规优化算法在求解有约束优化问题时惩罚系数难以选择、易陷入局部最优、所得优化解可能不满足约束等问题。将其应用于以产品质量为约束的精馏塔能耗优化问题,可以求解得到精馏塔系"卡边"生产对应的最优操作条件。仿真结果表明该建模和优化策略具有优化结果好、计算效率高等优点,可为工业精馏系统的操作优化提供有效支撑。     

基于遗传算法和广义交叉原理求解正则参数

本文研究了正则化方法中正则参数的求解问题,提出了一种新的正则参数求解策略,即利用遗传算法基于广义交叉检验准则求解正则参数,数值模拟验证了该方法的可行性和有效性。     

Systematic benchmarking of microarray data feature extraction and classification

A combination of microarrays with classification methods is a promising approach to supporting clinical management decisions in oncology. The aim of this paper is to systematically benchmark the role of classification models. Each classification model is a combination of one feature extraction method and one classification method. We consider four feature extraction methods and five classification methods, from which 20 classification models can be derived. The feature extraction methods are t-statistics, non-parametric Wilcoxon statistics, ad hoc signal-to-noise statistics, and principal component analysis (PCA), and the classification methods are Fisher linear discriminant analysis (FLDA), the support vector machine (SVM), the k nearest-neighbour classifier (kNN), diagonal linear discriminant analysis (DLDA), and diagonal quadratic discriminant analysis (DQDA). Twenty randomizations of each of three binary cancer classification problems derived from publicly available datasets are examined. PCA plus FLDA is found to be the optimal classification model.     

基于J48决策树算法的水质评价方法

提出一种以J48决策树算法为基础的水质评价方法。通过分析K最近邻法、贝叶斯分类法、向量空间模型法等常用的水质评价方法,针对某地区2008年的水质监测数据,采用J48决策树算法和单因子污染指数法对该地区水质进行计算,并对水质分类数据进行分类、属性清理和J48决策树参数的选择,以此建立模型。实验结果表明,该方法的交叉验证率约为95%,具有较好的分类评估效果。     

An Efficient Method To Estimate Bagging's Generalization Error

Bagging (Breiman, 1994a) is a technique that tries to improve a learning algorithm's performance by using bootstrap replicates of the training set (Efron & Tibshirani, 1993, Efron, 1979). The computational requirements for estimating the resultant generalization error on a test set by means of cross-validation are often prohibitive, for leave-one-out cross-validation one needs to train the underlying algorithm on the order of m times, where m is the size of the training set and is the number of replicates. This paper presents several techniques for estimating the generalization error of a bagged learning algorithm without invoking yet more training of the underlying learning algorithm (beyond that of the bagging itself), as is required by cross-validation-based estimation. These techniques all exploit the bias-variance decomposition (Geman, Bienenstock & Doursat, 1992, Wolpert, 1996). The best of our estimators also exploits stacking (Wolpert, 1992). In a set of experiments reported here, it was found to be more accurate than both the alternative cross-validation-based estimator of the bagged algorithm's error and the cross-validation-based estimator of the underlying algorithm's error. This improvement was particularly pronounced for small test sets. This suggests a novel justification for using bagging—more accurate estimation of the generalization error than is possible without bagging.     

CROSS-VALIDATORY CHOICE OF A SPECTRUM ESTIMATE AND ITS CONNECTIONS WITH AIC

Abstract. This paper concerns the use of a generalized version of the cross-validated log likelihood criterion (CVLL) for selecting a spectrum estimator from an arbitrary class of candidate estimators. It is shown that CVLL is asymptotically equivalent to the expected Kullback-Leibler information of the candidate estimator. The Akaike information criterion (AIC) is also asymptotically equivalent to Kullback-Leibler information, but the applicability of AIC is limited to parametric estimators. Thus CVLL can be viewed as a cross-validatory generalization of AIC. Monte Carlo results show that CVLL is able to provide an effective choice from a class of candidates which simultaneously includes autoregressive and classical smoothed periodogram estimators. To save computation time, CVLL can be evaluated only for the classical estimators while the computationally more efficient AIC is evaluated for the parametric estimators. The criterion values are all directly comparable in this case. As an additional computation-saving device, a non-cross-validatory version of CVLL for classical estimators is proposed and studied.