fisherclassify

matlab s函数（s-Function）编写的离散状态空间表达式(matlab s function (s-Function) written in discrete state space representation)

4) 创建一个名为Complex的类，进行复数的算术运算。复数的形式为： realpart+imaginarypart*i 用浮点变量表示类的private数据。提供构造函数，能够对所声明的该类对象进行初始化。在不提供初始化值的情况下，该构造函数应包含默认值。针对以下功能，分别提供一个public成员函数。 （1）两个复数相加； （2）两个复数相减 （3）用（a, b）的形式打印复数，其中a为实部，b为虚部。 (4) Create a class called Complex, and perform complex arithmetic operations. Plural form: realpart+imaginarypart* i said private data class with floating point variables. Provide a constructor, the object can be declared for the class is initialized. In the case do not provide initialization values, the constructor should contain default values. For the following functions, which provide a public member function. (A) two complex numbers (2) the subtraction of two complex (3) (a, b) in the form of plural printing, where a is the real part, b is the imaginary part.)

matlab支持向量机程序SVM_SteveGu(matlab support vector machine procedures SVM_SteveGu)

说明：   用最速下降法求最优化解 输入：f为函数名 grad为梯度函数 x0为解的初值 TolX,TolFun分别为变量和函数的误差阈值 dist0为初始步长 MaxIter为最大迭代次数 输出: xo为取最小值的点 fo为最小的函数值 f0 = f(x(0))( Steepest Descent Method with Optimum Solution input: f as a function name grad is gradient function x0 for the solution of the initial TolX, TolFun variables and functions were error threshold dist0 as the initial step MaxIter maximum Diego passage number Output: xo to take the minimum point of fo is the smallest function value f0 = f (x (0)))

最小均方算法（MSE），来自阵列信号处理课本中(Least mean square (MSE), from the array signal processing textbooks)

用香农编码计算概率分布，信息论与编码实验内容。(The Shannon coding probability distribution, information theory and coding experiment content.)

灰色模型算法，通过改进之后，考虑了之后的误差处理，减少了因为数据变化太大而引起的预测不准确的现象。(Gray model algorithm, improvements, after taking into account the error handling, reduced because the data has changed so much caused by inaccurate forecasts phenomenon.)

仿真效率很高的，包含特征值与特征向量的提取、训练样本以及最后的识别，直线阵采用切比学夫加权控制主旁瓣比，包括广义互相关函数GCC时延估计，含噪脉冲信号进行相关检测。(High simulation efficiency, Contains the eigenvalue and eigenvector extraction, the training sample, and the final recognition, Linear array using cut than learning laid upon the right control of the main sidelobe ratio, Including the generalized cross-correlation function GCC time delay estimation, Noisy pulse correlation detection signal.)

仿生优化计算方法，可适用于基于选择类问题求解(Metaheuristic method, which can be used for problems based the requirement of sample selection)

最优化算法 PSO.m POWELL.m 互信息尺度 MI.m EMI.m RMI.m HiMI.m GMI.m FPMI.m(Optimization algorithm PSO.m POWELL.m-scale mutual information MI.m EMI.m RMI.m HiMI.m GMI.m FPMI.m)