fuzzy_sugeno

用超球面支持向量机进行多类分类,每个超球面覆盖一个类(Hypersphere support vector machines using multi-class classification)

下面是5 MATLAB文件。 1。readsounds。米,用于读取数据 2。sep.m,一个学习通过数据的代码 3。sepout。米,可选的文本输出 4。wchange。米,体重变化的大小和方向 5。sep.run,2 - > 2分离的示例脚本(Below are 5 MATLAB files. 1. readsounds.m, for reading the data in 2. sep.m, the code for one learning pass thru the data 3. sepout.m, for optional text output 4. wchange.m, tracks size and direction of weight changes 5. sep.run, an example script for 2->2 separation )

粗糙表面 G-W模型的接触刚度仿真程序(Rough surface Contact stiffness simulation program for G-W model)

image compression using wavelets

频域反馈均衡器仿真设计，采用FFT快速变换，抗通信中的多径干扰(frequency decision feedback equaliter，use FFT，in order to estimate the ISI in communication)

When doing optimization, GA algorithm was initially selected, but the instability of the GA (or local optimization) is really frustrating, before and after the resulting difference can at times reach more than 30 , at that time due to time reasons, I chose to optimize 1000 times, then the minimum value of the most optimal solution. Although the problems are solved, but an academic, this style of running is pretty embarrassing. So, I downloaded a GA-PSO algorithm, try using the combination of GA and PSO optimizing strategies, results algorithm issues, efficiency and good. I downloaded the original algorithm, there is a problem is that it is aimed at all upper and lower limits of the design variables are the same, so I modified the program and improvements, you can now deal with upper and lower limits of the inconsistency problem and fix the bug. Now, program modifications, and share, hoping to be useful to you

MATLAB 电机仿真教程，非常详细的仿真，希望对大家有帮助。(MATLAB motor simulation tutorial, a very detailed simulation, we want to help.)

利用SVM做回归线性测试。对于大盘指数的有效预测可以从整体上观测股市的变化提供强有力的信息，所以对上证指数的预测很有意义。通过对上证指数从1990.12.20-2009.08.19每日的开盘数进行回归分析，最终拟合的结果是：均方误差MSE=2.35705 e-005，平方相关系数 R=99.9195 。SVM的拟合结果还是比较理想的。(Make use of SVM linear regression testing. For the market index can predict the observed changes in the stock market as a whole to provide a strong information, so it makes sense to predict the Shanghai index. Through the Shanghai Composite Index opened daily from 1990.12.20-2009.08.19 number of regression analysis, the results of the final fit are: mean square error MSE = 2.35705 e-005, the square of the correlation coefficient R = 99.9195 . SVM fitting result is quite satisfactory.)

说明：  kmeans聚类分析，无监督学习实现Matlab代码(Kmeans clustering analysis, unsupervised learning implementation of MATLAB code)

用matlab 实现 对图像去其高频分量,再反变换 内含文本文档 输入 MATLAB 后可直接运行(Image with matlab implementation to its high frequency, and then inverse transform)