advshinners

说明：  该课件作为理工科各专业高年级本科生、研究生学习小波理论与应用的参考资料，对于小波的认识有很大的帮助！(wavelet)

采用信号包络和差分方式进行信号模式识别，通过变换产生时域解析信号SKH ，即信号的复包络，得到零中心归一化瞬时绝对值的标准偏差。 (Differential use of signal envelope and the way the signal pattern recognition, transformation generated by time domain analytic signal SKH, the signal complex envelope, zero-center normalized standard deviation of the instantaneous absolute value.)

有用的SGA的Matlab源代码，求函数的最大值，根据SGA的C代码改变，很容易懂，对SGA 学习很有帮助(useful SGA Matlab source code, the maximum demand function, According to the SGA C code change, it is easy to understand and very helpful to SGA)

Reconocimiento de patrones de voz.

fire using C++ //globals LPVGA lpVga=NULL //pointer to CVga LPPALETTE lpPalette=NULL //pointer to palette

说明：  蒙特卡洛计算的程序，可用于电力系统中的潮流优化，并且可获得很好的结果。(The program of Monte Carlo calculation can be used for power flow optimization in power system, and good results can be obtained.)

Quade test for non parametric two way ANalysis Of VAriance

Pattern Recognition and Machine Learning Solutions to the Exercises

：滤波 fs=400 采样频率,必须大于两倍基带信号最高频率 ts=1/fs 采样时间间隔 T=5 时间窗大小 t=0:ts:T (hf1=zeros(1,N) hf1(f<=B1&f>=-B1)=1 y1=hf1.*fftshift(y) figure subplot(211) plot(f,abs(fftshift(y))/N,f,hf1, r-- ) axis([-fs/2,fs/2,0,1.2]) title( 原信号频谱 ) xlabel( f /Hz ) subplot(212) plot(f,abs(y1)/N) title( 低通滤波，截止频率90Hz ) xlabel( f /Hz ))

改进boll谱减法，语音增强效果有提高。与MMSE算法不相上下(Improved the boll spectral subtraction speech enhancement effect has been improved. Comparable with the MMSE algorithm)