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...

自己写的一个简单的神经网络(BP)算法，用来解决异或（XOR)问题,初学神经网络的人可以下来看！-Wrote it myself a simple neural network (BP) algorithm, to solve the exclusive-OR (XOR) problem, learning neural network can look at the next!

K-均值聚类算法的编程实现。包括逐点聚类和批处理聚类。K－均值聚类的的时间复杂度是n*k*m，其中n为样本数，k为类别数，m为样本维数。这个时间复杂度是相当客观的。因为如果用每秒10亿次的计算机对50个样本采用穷举法分两类，寻找最优，列举一遍约66.7天，分成3类，则要约3500万年。针对算法局部最优的缺点，本人正在编制模拟退火程序进行改进。希望及早奉给大家，倾听高手教诲。-K-means clustering algorithm programming. Point by point, including clustering and clustering batch. K-means clustering of the time complexity of n* k* m, n samples, several types of k, m sample dimension. The time complexity is a very objective. Because if we use one billion times per second the computer using 50 samples of two exhaustive method to find the optimal set out again about 66.7 days, divided into three categories, offering 3,500 years. Local optimal algorithm against the shortcomings, I was prepared simulated annealing process improvements. Early Feng hope for everyone, listen to the master teachings.

模拟退火算法用于求解旅行商问题的matlab源程序-simulated annealing algorithm for solving the traveling salesman problem Matlab source

matlab所有的工具箱，包括统计工具箱、拟合工具箱、微分方程工具箱、遗传算法工具箱、模拟退火工具箱、优化工具箱。随机数工具箱、等等。很全面-All matlab toolbox, including statistical toolbox, fitting toolbox, toolbox differential equations, genetic algorithm toolbox, simulated annealing Toolbox, Optimization Toolbox. Random number toolbox, and so on. Very comprehensive

微分进化算法的Labview版本，内为一个无约束连续变量的全局优化问题-Differential evolution algorithm Labview version, for a continuous variable unconstrained global optimization problem

支持向量机(4)mySVM-Support Vector Machines (4) mysvm

模式识别的基础－－相似度量的介绍，对大家可能有帮助-based pattern recognition-- similar to the measure, we might have to help

Counterpropagation Network Vision 源码, 经典的CPN人工神经网络例子源码-Counterpropagation Network Vision Source, a classic example of artificial neural network CPN-source

细胞自动机的，很好的-Cellular Automata and fine

he algorithm is equivalent to Infomax by Bell and Sejnowski 1995 [1] using a maximum likelihood formulation. No noise is assumed and the number of observations must equal the number of sources. The BFGS method [2] is used for optimization. The number of independent components are calculated using Bayes Information Criterion [3] (BIC), with PCA for dimension reduction.