evolution of neural network topology structure of the mine

本课题基于汉语普通话客观质量评价的模型、利用MOS主观评价方法和客观评价方法对语音样本进行评价。并分析客观评价方法的优缺点及在研究中遇到的问题和改进方法，其主要研究内容是利用MATLAB的建立语音客观评价模型，涉及到人工神经网络，数字信号处理等方面的内容。-The issue of objective quality evaluation based on the Mandarin model, using MOS subjective evaluation method and objective evaluation of speech samples evaluated. And analyze the advantages and disadvantages of an objective evaluation method of the problems encountered in the study and improvement of methods, the main research content is to use MATLAB model for the establishment of objective evaluation of speech, involving artificial neural networks, digital signal processing and other aspects.

以黎曼几何为理论依据，基于S.Amari的修正核函数思想提出了两种新的保角变换，用其对核函数进行数据依赖性改进，进一步提高支持向量机分类器泛化能力。以人工非线性分类问题 为对象进行研究，仿真实验结果表明采用新保角映射可以快速显著地改善分类器泛化性能，而且能大幅度地减少支持向量的数目。-Two novel conformal transformations were proposed based on the Riemannian geometry theory and S.Amari’sidea.And the kernel function was modified by the transformation in a data-dependent way.Our experimental results for theartificial nonlinear data set show that the generalization performance of support vector machines classifier is improvedremarkably and the number of support vectors is decreased greatly.

编写的BP算法的软件模拟 在Turbo c++3.0环境下编译通过-prepared by BP algorithm in the software simulation 3.0 Turbo c compiler environment through

基本遗传算法 里面的函数可自行更换 适合初学者 容易理解-A function of the basic genetic algorithm which can be suitable for beginners is easy to understand on their own replacement

神经网络算法 神经网络算法 -BP networksBP networksBP networksBP networksBP networks

它是一个基于遗传算法的优化方法，使用微分运算符。它用于解决带有大量未知事物的优化问题。-it is based on a genetic algorithm optimization method, the use of Differential Operators. It used to solve a large number of unknown things with the optimization problem.

这个程序是人工智能里字句归结的一个程序，它采用了线性归结策略和删除策略相结合的方法-the procedure is artificial intelligence Lane words attributed to the procedure, which uses the linear sum strategies and strategy to delete a combination of methods

该算法为模糊自适应整定的PID控制算法，该算法在matlab环境下开发，可以直接运行 -The algorithm for fuzzy adaptive tuning of the PID control algorithm in the matlab environment development, can be directly run

将遗传算法(GA)与传统SVM算法结合,构造出一种参数最优的进化SVM(GA2SVM),SVM 模型采用径向基函数(RBF)作为核函数，利用格雷码编码方式对SVM算法的模型参数进行遗传编码和优化搜索，将搜索到的优化结果作为SVM 的最终模型参数。-Genetic algorithm (GA) combined with the traditional SVM algorithm, a kind of tectonic evolution of the optimal parameters of SVM (GA2SVM), SVM model using Radial Basis Function (RBF) as kernel function, the use of Gray code encoding algorithm of the SVM model parameters of genetic coding and optimization of search, will search for the optimal results as the final SVM model parameters.

自组织 Kohonen 映射程序，当一个神经网络接受外界输入模式时，将会分为不同的区域，各区域对输入模式具有不同的响应特征，同时这一过程是自动完成的。各神经元的连接权值具有一定的分布。最邻近的神经元互相刺激，而较远的神经元则相互抑制，更远一些的则具有较弱的刺激作用。自组织特征映射法是一种无教师的聚类方法。 -Kohonen self-organizing map process, when a neural network to outside input mode, will be divided into different regions, the regional input to the model with different response characteristics and the process is done automatically. The neurons connect with the right to a certain value of the distribution. Most neighboring neurons stimulate each other, distant neurons were mutual inhibition, the vision has a weaker stimulus. Self-organizing feature mapping method is a non-teachers clustering method.