GBMM1D

模式分类的Matlab工具箱，各种实用算法(Pattern classification of the Matlab toolbox, various practical algorithm)

反馈网络是神经网络的基础 这里提供了神经网络中用到的反馈 这里能充分理解文档的内容就能很好的掌握神经网络的基础。我将各个函数写成了不同的文档 就是便于大家理解。(Feedback network is the basis of the neural network here provides feedback neural network is used here to fully understand the content of the document can very good master the basis of neural network. I will each different function in the form of a document is easy to understand.)

几种分类器，包括adaboost和随机森林(Several classifiers including adaboost and random forest)

包含一元线性回归分析、多元线性回归分析、逐步回归分析、非线性回归分析、主成分分析、因子分析、层次聚类分析、判别分析、自相关分析和自回归分析的全套Matlab源代码(MATLAB source code including linear regression analysis, multiple linear regression analysis, stepwise regression analysis, non-linear regression analysis, principal component analysis, factor analysis, hierarchical clustering analysis, discriminant analysis, autocorrelation analysis and autoregressive analysis)

说明：  与连续信号的解调相比，短时突发信号持续时间短，不适用性能良好的反馈环路算法，需要使用前向幵环算法。与一般突发信号相比，短时突发信号没有训练序列、同步头等辅助信息，所有参数需要利用有限的数据估计。与合作通信相比，盲接收信号参数的初始偏差较大，估计范围小的算法不适用。每个突发包的数据有限，解调过程不允许丢失任何符号。(Compared with the demodulation of continuous signal, short burst signal has a shorter duration and is not suitable for the feedback loop algorithm with good performance. Forward loop algorithm is needed.)

SVD去噪，构造的矩阵方法为对数据进行均匀分段，并对去噪前后的信噪比，均方差进行了对比。(SVD denoising, construction of the matrix method for the data sub-uniformity, and signal to noise ratio before and after denoising are compared variance.)

This code computes fft and ifft

多尺度特征熵，用于机械振动信号特征提取和故障诊断等方面(Multi-scale feature entropy for mechanical vibration signal feature extraction and fault diagnosis)

说明：  种子优化算法（SOA）matlab源程序，测试Rosenbrock等三个函数，结果很好，可以借鉴学习(Seed optimization algorithm (SOA) matlab source code, testing three functions, the results very good, can learn to learn)

matlab经典的100个实验程序。非常适合初学者学习。(matlab classic 100 experimental procedure. Very suitable for beginners to learn.)