《第四讲随机变量函数与多维正态随机变量.ppt》由会员分享,可在线阅读,更多相关《第四讲随机变量函数与多维正态随机变量.ppt(59页珍藏版)》请在三一办公上搜索。
1、本次课内容随机变量函数 单变量的函数 多变量的函数 随机变量函数的数字特征多维正态随机变量,1.随机变量变换后概率密度的确定,例:设Y=aX+b(Affine Transformation),求Y的概率密度。,如果X服从正态分布,,A linear function of a Gaussian random variable is also a Gaussian random variable.,如果不是单调函数,假定有n个反函数,则,Example:Square-Law device,b0,If X is a Gaussian random variable with mean zero a
2、nd variance 2,Example:,This PDF is called the log-normal PDF,2.Function of a Discrete Random Variable,How to determine the PMF of a transformed random variable?,For example,toss a die and transform as follow,In general,we have that,Check Yourself,Which of following is correct?,One-to-One:,ABCDE,None
3、 of above,Check Yourself,Which of following is correct?,One-to-One:,ABCDE,None of above,Check Yourself,Suppose Y=2X,Which of following are correct?,None of above,ABCDE,Check Yourself,Suppose Y=2X,Which of following are correct?,None of above,ABCDE,对于一对一的变换,,Example:Many-to-One Transformation,Example
4、:,3.Continuous-to-Discrete,Example:PDF for amplitude-limited Rayleigh RV.,4.Mixed random variable,Assume,一般说来,如果g(x)在区间(x0,x1上为一常数A,则FY(y)在y=A处不连续,跳变点跳变高度为,fY(y)在y=A处有一函数,函数的强度为,例,当y时,当y时,当y时,FY(y)在y=0处有一个跳变,跳变的高度为P-cXc=FX(c)-FX(-c),Example:Hard Limiter,Y为离散型随机变量,FY(y)为阶梯型,1,-1,5.多维随机变量的函数,Jacco 比行列
5、式,例,当X1,X2相互独立时,,如果Xi服从相同分布,多个函数卷积的结果逐渐趋近一个高斯函数,所以大量独立同分布的随机变量之和服从正态分布(中心极限定理)。,6.Expected value for a function of a random variable,均值:,方差:,均值:,方差:,对于离散型随机变量,也有类似的结果:,Example:Square-LawLet X be a noise voltage that is uniformly distributed in Sx=-3,-1,1,3 with pX(k)=1/4.Find E(Y)where Y=X2.,The Fir
6、st Approach:find the PMF of Y,The Second Approach:,Example:Expected value of a sinusoidal with random phase,随机变量的函数概率分布和概率密度的 确定离散随机变量的函数 One-to-One Transformation Many-to-One Transformation连续-离散的变换混合型随机变量多维随机变量的函数,学习内容:正态随机变量二维正态随机变量多维正态随机变量正态随机变量的线性变换,教学目的:掌握多维正态随机变量的理论,为学习正态随机过程打下基础。,多维正态随机变量,1.N
7、ormal Random Variable(Gaussian RV),The sum of large number of iid random variables is distributed to normal PDF.,CDF,标准正态分布,If X has a Rayleigh density,2.Joint Normal random variable,如果r=0不相关,X1与X2是独立的,对于正态随机变量而言,不相关与独立是等价的。,特征函数:,为了简化表达式,概率密度可用矩阵形式表示,条件分布:,3.多维正态随机变量的分布,特征函数,不相关也必定独立,如果不相关,4.正态随机变量的线性变换,正态随机矢量经过线性变换后仍为正态的。,混合矩:,的均值为零,