基于梯度算法的支持向量机参数优化方法

首先介绍最近出现的参数优化方法,概括了高效率的参数优化算法应具备的若干特点 .然后提出了一种新的支持向量机参数优化方法.该方法先在局域内用混沌优化搜索局域最优点,再将此最优点作为梯度方向,通过改变局域范围跳出局部寻优区域. 该方法降低了对性能函数连续且可微的要求,收敛速度快,最终优化解与支持向量机的参数初始值无关.最后,通过仿真实验表明了该方法具有更高的分类和回归准确率.

     

采用遗传算法优化最小二乘支持向量机参数的方法

支持向量机是建立在统计学习理论上的一种学习算法,较好地解决了小样本学习问题.由不同的参数和核函数构造的支持向量机在性能上存在很大差异,而在参数和核函数的选择上目前还没有明确的理论依据.针对支持向量机的参数选择问题,提出了一种采用遗传算法优化最小二乘支持向量机参数的方法.结合LS-SVMlab工具箱,在MATLAB实验平台的仿真实验表明,该方法提高了支持向量机的参数选择效率,得到的参数对测试样本的分类结果是最优的,从而避免了人为设定参数的不足,同时缩短了优化时间.     

IA-SVM算法在网络入侵检测中的研究

研究网络入侵检测问题,网络入侵具有不确定性、多变性和动态性,传统检测方法不能很好的识别这种特性,且传统支持向量机参数采优化方法易出现参数选择不当,导致网络入侵检测准确率低.为了提高网络入侵检测准确率,将免疫算法引入到网络入侵检测中,用其优化支持向量机参数.方法将网络入侵检测数据输入到支持向量机中学习,将支持向量机参数作为免疫算法的抗体,把网络入侵检测准确率作为免疫算法抗原,通过抗体和抗原相互作用得到最优的支持向量机参数,然后对网络入侵数据检测得到入侵检测结果,最后通过DRAP网络入侵数据集对该方法进行仿真.仿真结果表明,相对传统网络入侵检测方法,新方法学习速度快,检测准确率高,很好地解决了传统检测方法准确率低的难题,为网络安全提供了保障.     

基于改进PSO优化LSSVM的传感器补偿研究

针对最小二乘支持向量机在对传感器进行补偿时,正则化参数和核函数参数对补偿精度影响较大的问题,提出一种利用改进的粒子群优化算法优化最小二乘支持向量机模型参数的传感器补偿方法。该方法利用改进的粒子群优化算法优化最小二乘支持向量机模型的正则化参数和核函数参数,避免了人工选择参数的盲目性,提高了最小二乘支持向量机模型的预测精度。仿真实验表明,在传感器的补偿时,该方法比最小二乘支持向量机模型的补偿精度更高。     

遗传支持向量机模型优化在入侵检测中的应用

针对目前入侵检测检测精度低的问题,根据遗传和支持向量机算法的特点,建立了一种遗传支持向量机模型。该模型首先用遗传算法优化支持向量机参数,再用优化后的支持向量机构建入侵检测模型,使用该模型进行入侵检测。实验通过讨论了支持向量机参数的选择对检测精度的影响,选取了合适的参数(c,σ)。结果表明,把这种遗传支持向量机模型用于入侵检测提高了检测精度。     

支持向量机的参数优化及其文本分类中的应用

支持向量机的优化算法对准确检索所需信息资料很重要.传统支持向量机参数寻优方法速度慢、运算量大,具有一定的盲目性.针对准确快速检索到所需信息,为提高支持向量机算法的性能,提出了一种采用免疫算法对支持向量机参数进行优化的文本分类方法(IA-SVM).将支持向量机模型参数作为抗体的基因设计了抗体的编码方案,利用人工免疫算法对支持向量机的惩罚因子和径向基核函数进行优化搜索,使支持向量机的分类性能最优.实验结果表明,IA-SVM算法减少了对支持向量机参数选择的盲目性,在文本分类问题上明显提高了分类正确率和检索速度.     

基于蚁群算法的支持向量机参数选择方法研究

研究支持向量参数选择优化问题,常用的支持向量机参数优化算法和遗传算法分别存在耗时长和易陷入局部最优值的缺陷,导致支持向量机的分类精度低.为了解决支持向量机参数优化问题,提出了基于蚁群算法的SVM分类器泛化方法.蚁群算法是一种优化搜索方法,具有较强的鲁棒性、优良的分布式计算机制,SVM参数的选取看作参数的组合优化,建立组...     

基于支持向量回归机的HHT边界效应处理

针对希尔伯特-黄变换中的边界效应，提出了基于支持向量回归机的时间序列预测方法．在支持向量回归机的应用当中，参数的选取对它的泛化性能有很大影响．在讨论了参数对支持向量回归机的泛化性能的影响基础上，提出了通过微粒群优化算法来优化支持向量回归机参数的方法，使得支持向量回归机在应用中能够自适应的选择最优参数，从而获得了更好的泛化性能，提高了在端点处的延拓精度，很好地抑制了端点效应．试验表明，该优化算法能够很好解决支持向量回归机的参数选取问题．通过与神经网络的延拓方法和黄等人的HHTDPS结果对比，基于支持向量回归机的时间序列预测方法可以更好地解决在希尔伯特-黄变换中存在的边界效应，得到的固有模态函数具有较小的失真．     

基于粒子群算法的煤气化过程LS_SVM预测模型

针对最小二乘支持向量机参数选择对模型性能的重要影响,并且以往的参数优选方法效果差且耗时长这一问题,提出基于粒子群算法优化最小二乘支持向量机预测模型.该模型用最小二乘支持向量机理论建立,用粒子群算法优化模型参数.论文将此模型用于预测评价固定床煤气化气化效果的三个主要性能指标(气体热值、气化效率、气体产率),通过现场实际数据仿真结果表明,该算法有效地提高了模型预测精度,验证了此模型的可靠性和可用性.     

基于改进遗传算法的支持向量机参数优化

支持向量机是一种非常有前景的学习机器,但是,支持向量机参数的选取一直没有一套成熟的理论,这给支持向量机的应用带来了很大的不便.为此,本文提出了基于改进遗传算法的支持向量机的参数优化方法,利用遗传算法的全局搜索能力得到支持向量机的最优参数值.仿真实验结果表明,得到的参数可使支持向量机具有良好的泛化性能,此方法切实有效.     

Implicit function theorem for nonsmooth systems by means of exhausters

The implicit function theorem for nonsmooth systems is formulated by means of new tools of nonsmooth analysis — exhausters. Exhausters exist for any directionally differentiable function whose directional derivatives are directionally continuous functions. It is possible to find an implicit function with the help of exhausters for functions that are not Lipschitz and not quasidifferentiable but directionally differentiable. Examples of calculating nonsmooth implicit functions are given.     

A Class of Continuously Differentiable Filled Functions for Global Optimization

The filled function method is an approach to find global minima of multidimensional multimodal functions. This paper proposes a class of new filled functions that are continuously differentiable and do not include exponential terms. The performance of the new function in numerical experiments for a large set of testing functions up to 40 dimensions is quite satisfactory.     

Training of a discrete recurrent neural network for sequence classification by using a helper FNN

This research is concerned with a gradient descent training algorithm for a target network that makes use of a helper feed-forward network (FFN) to represent the cost function required for training the target network. A helper FFN is trained because the cost relation for the target is not differentiable. The transfer function of the trained helper FFN provides a differentiable cost function of the parameter vector for the target network allowing gradient search methods for finding the optimum values of the parameters. The method is applied to the training of discrete recurrent networks (DRNNs) that are used as a tool for classification of temporal sequences of characters from some alphabet and identification of a finite state machine (FSM) that may have produced all the sequences. Classification of sequences that are input to the DRNN is based on the terminal state of the network after the last element in the input sequence has been processed. If the DRNN is to be used for classifying sequences the terminal states for class 0 sequences must be distinct from the terminal states for class 1 sequences. The cost value to be used in training must therefore be a function of this disjointedness and no more. The outcome of this is a cost relationship that is not continuous but discrete and therefore derivative free methods have to be used or alternatively the method suggested in this paper. In the latter case the transform function of the helper FFN that is trained using the cost function is a differentiable function that can be used in the training of the DRNN using gradient descent.Acknowledgement. This work was supported by a discovery grant from the Government of Canada. The comments made by the reviewers are also greatly appreciated and have proven to be quite useful.     

A new filled function method for unconstrained global optimization

Many real world problems can be modelled as optimization problems. However, the traditional algorithms for these problems often encounter the problem of being trapped in local minima. The filled function method is an effective approach to tackle this kind of problems. However the existing filled functions have the disadvantages of discontinuity, non-differentiability or sensitivity to parameters which limit their efficiency. In this paper, we proposed a new filled function which is continuous and differentiable without any parameter to tune. Compared to discontinuous or non-differentiable filled functions, the continuous and differentiable filled function mainly has three advantages: firstly, it is not easier to produce extra local minima, secondly, more efficient local search algorithms using gradient information can be applied and thirdly, a continuous and differentiable filled function can be optimized more easily. Based on the new proposed filled function, a new algorithm was designed for unconstrained global optimization problems. Numerical experiments were conducted and the results show the proposed algorithm was more efficient.     

基于粗神经网络的仿人智能机器人的语音融合算法研究

将粗集合理论与神经网络相结合，提出一种基于粗神经网络的新的信息融合方法，用于仿人智能机器人的语音融合。该方法不仅可以接受定量输入，而且可以接受定性输入，即输入是一个范围，或在观测时间内输入是变化的。由于粗神经网络的误差传递函数不可微，所以采用遗传算法来训练粗神经网络。仿真实验结果表明，基于粗神经网络的信息融合方法有效地提高了语音的识别率。     

On Discrete Minimax Problems in R Using Interval Arithmetic

Interval algorithms for bounding discrete minimax function values of problems in which the constituent minimax functions are continuously differentiable functions of one real variable in a bounded closed interval are presented, both with and without inequality constraints represented by continuously differentiable constraint functions.     

An elementary noniterative quadrature-type method for the numerical solution of a nonlinear equation

A simple noniterative method for the numerical determination of one simple root of a nonlinear differentiable algebraic or transcendental function along a finite real interval is proposed. This method is based on the computation of an integral involving the above function both by the Gauss- and the Lobatto-Chebyshev quadrature rules for regular integrals and equating the obtained results. The convergence of the method is proved under mild assumptions and numerical results for two classical transcendental equations are presented.     

Interpolation of fuzzy data by Hermite polynomial

We consider the interpolation of fuzzy data by a differentiable fuzzy-valued function. We do it by setting some conditions on interpolant and first derivative of the interpolant. We give a numerical method for calculating this function.     

一种融合程序员和神经网络的自动化程序生成方法

程序生成是人工智能的核心研究问题之一,当前输入输出样例驱动的神经网络模型是非常流行的研究方法.面临的主要挑战是泛化能力差、生成程序准确率保证、难以处理复杂程序结构(如分支、循环、递归等),主要原因是模型的输入信息单一(输入输出对)和完全依赖神经网络.显然单一地通过输入输出样例倒推程序行为存在歧义性,而神经网络的记忆容量很难满足常规程序的变量存储需求.提出一种人工与神经网络生成相协作的编程模型,融合神经网络和程序员各自的优势,其中程序员用高级编程语法编写程序框架,神经网络自动学习生成程序局部的琐碎细节,从而促进自动化程序生成方法更好地应对实际应用挑战.实验表明,研究方法是有效的,跟同类代表性研究方法相比表现出更好的学习性能.     

用Fortran和Excel的联合方法绘制科技图线

科技中的数值计算和数值模拟,用Fortran编程获得数值结果有着突出的优点,Microsoft Excel在按数据绘制图线方面有着丰富的功能.将两者联合高效地获得了图示结果,是一种值得重视的方法.尤其在按一份数据文件绘制几张图,每张图又有几条曲线时,此法的简捷性尤为突出.对此,阐明了一些有效的建议,并给出了无处可微连续函数的图示结果.