EMD_MAT

基于变换编码的图像压缩及解压，含有离散余弦变换程序。(Transform based image compression and decompression coding)

利用此程序，可以对图片的直方图特征进行提取，进而进行图片的分类，并统计出分类的正确率。(With this program, you can picture the extraction, and then classify pictures and statistics of the correct rate of classification.)

基于支持回归向量机的端点延拓程序解决EMD端点效应，包括向左延拓和向后延拓。(EMD solve end effect based on support vector machine regression endpoint extension programs, including the left rear extension and extension.)

说明：  自己写的关于用matlab画的bezier curve的程序，可以在次基础上，添加曲线修补的功能(Ploting the bezier)

Fast Earth Mover s Distance similarity matrix calcutation

最大似然译码 最大似然译码 最大似然译码(Maximum likelihood decoding of maximum likelihood decoding of maximum likelihood decoding)

用于去除背景，得到原图与背景相减的图像并二值化(To remove the background, the subtraction image and background image and binarization)

matlab gui界面主函数，所有界面函数的基础(matlab gui interface, the main function, the basis of all interface functions)

基于最小二乘支持向量机理论,结合某风电场实测风速数据,建立了最小二乘支持向量机风速预测模型。对该风电场的风速进行了提前1h的预测,其预测的平均绝对百分比误差仅为8.55 ,预测效果比较理想。同时将文中的风速预测模型与神经网络理论、支持向量机(support vector machine,SVM)理论建立的风速预测模型进行了比较。仿真结果表明,文中所提模型在预测精度和运算速度上皆优于其他模型。 (Based on least squares support vector machine theory, combined with a wind farm measured wind speed data, the establishment of a wind vector machine forecasting model of least squares support. The velocity of the wind farm were predicted in advance 1h, the mean absolute percentage error of only 8.55 of its forecast, forecast effect is ideal. While the text of the wind speed forecasting models and neural networks, support vector machines (support vector machine, SVM) wind speed prediction models were compared with established theories. Simulation results show that our proposed model on prediction accuracy and computing speed are superior to other models.)

进行逐步线性回归，阐述了负荷预测的应用研究,通过虚拟阵元进行DOA估计，最小均方误差（MMSE）的算法，添加噪声处理，相控阵天线的方向图（切比雪夫加权），直线阵采用切比学夫加权控制主旁瓣比。( Stepwise linear regression, It describes the application of load forecasting, Conducted through virtual array DOA estimation, Minimum mean square error (MMSE) algorithm, Add noise processing, Phased array antenna pattern (Chebyshev weights), Linear array using cut than learning laid upon the right control of the main sidelobe ratio.)