filter

Recursive Least Squared method is implemented this time by adding noise to input. This program objective is to identified parameters of third order transfer function.

解决了非线性函数的逼近问题,实现了BP神经网络在非线性函数逼近中的应用。得出了信号的频率与隐含点之间，隐含点与函数逼近能力之间的关系——隐含层神经元数目越多。(Solved the problem of nonlinear function approximation, and realized the application of the BP neural network in nonlinear function approximation. Between signal frequency and implicit points are drawn, implied the relationship between the point and the function approximation ability, the more number of hidden layer neurons. )

把HHT和神经网络结合起来，识别出是否故障。对于HHT变换，这里用到的是它的IMF分解，然后用能量理论来判别哪些模组是虚假分量，哪些是是真实分量。对于EMD分解后的谱进行特征提取 ，利用的理论基础就是模糊熵。计算出真实 分量的模糊熵，作为输入层；输出层为两个神经元，为0（故障 ）1（正常 （正常 ）判别该信号是否出现故障。(The combination of HHT and neural network to identify whether the fault. For the HHT transform, here is the use of its IMF decomposition, and then use the energy theory to determine which module is false, which is true component. The theoretical basis for the feature extraction of the spectrum after EMD decomposition is fuzzy entropy. The fuzzy entropy of the real component is calculated, as the input layer, the output layer is two neurons, which is 0 (fault) 1 (normal) to determine whether the signal is faulty.)

基于svm的数据分类预测，数据集是意大利葡萄酒种类的数据集，对葡萄酒进行种类识别以及分类。(Based on the svm data classification prediction, the data set is the Italian wine category data set, the wine species identification and classification.)

Harris_conner_detection，用MATLAB实现Harris角点检测算子的源码，用于特征点提取及图像配准等。(Harris_conner_detection, use MATLAB to achieve Harris corner detection operator of the source code for the feature point extraction and image registration and so on.)

边缘检测算子laplacian算子 matlab源码(laplacian matlab)

对仿真信号进行特征值分解，进行奇异值降噪处理(SVD Denoise)

simulated of throw disk

说明：  可以降维分析，还有bp神经网络，非常方便学习。大家可以随意下载，(Dimensionality reduction analysis)

说明：  fdtd 平面波的加入 计算建模 先上传这个，下边的以后补齐(fdtd calculation of plane-wave model to join the first upload of this, after the bottom of the padded)