kalman-filter-cs

根据牛顿插值原理进行数据的拟合，附有牛顿插值原理和它的matlab的实现方法，有点简单了，但是过程中的好些地方还是值得学习的。(Newton interpolation carried out in accordance with the principle of data fitting, with Newton interpolation matlab principle and its realization method, a bit simple, but in the process of a number of places it is still worth learning.)

说明：  极限学习机是一种神经网络仿真，比BP,SWM学习速度更快，参数设置更容易(Extreme Learning Machine is a kind of neural network simulation, compared with BP, SWM learning faster, easier to set parameters)

最新粒子群matlab算法，适用性强，傻瓜式套用(New particle swarm algorithm matlab, applicability, Fool Apply)

使用支持向量机进行特征选择的集成算法模型(Feature selection of support vector machines)

神经网络中的对象传播神经网络CPN的学习算法(Neural network objects CPN-propagation neural network learning algorithm)

说明：  这是一个混沌理论中重要的Logistic映射的程序。通过曲线很容易看到奇妙的混沌现象(This is a chaos theory important Logistic mapping procedures. Through curve is easy to see the amazing phenomenon of Chaos)

应用sift算子进行图像匹配，效果不错。。。(sift is used to match two picture，the result is good...)

基于SVR的线性系统故障诊断，假设线性系统在1.4s~1.6s期间发生前向通道恒增益故障，恒增益变化系数为1.8(Linear systems based on SVR fault diagnosis, assuming a linear system in 1.4s ~ 1.6s during forward channel constant gain fault, constant gain coefficient of variation of 1.8)

基于tdoa定位的chan算法，经典算法，基本实现(Based tdoa positioning chan algorithm, the classical algorithm, the basic realization)

画中画检测，实现了对视频画中画和黑框的检测。(Picture in picture detection. )