定义 
假设X 为一n  × p 矩阵,其各行(row)来自同一均值向量为
  
    
      
        
          0 
         
       
     
    {\displaystyle \mathbf {0} } 
      
  
    
      
        p 
       
     
    {\displaystyle p} 
      多变量正态分布 且彼此独立 。
  
    
      
        
          X 
          
            ( 
            i 
            ) 
           
         
        
          = 
         
        ( 
        
          x 
          
            i 
           
          
            1 
           
         
        , 
        … 
        , 
        
          x 
          
            i 
           
          
            p 
           
         
        
          ) 
          
            T 
           
         
        ∼ 
        
          N 
          
            p 
           
         
        ( 
        0 
        , 
        V 
        ) 
        , 
       
     
    {\displaystyle X_{(i)}{=}(x_{i}^{1},\dots ,x_{i}^{p})^{T}\sim N_{p}(0,V),} 
      则威沙特分布为
  
    
      
        p 
        × 
        p 
       
     
    {\displaystyle p\times p} 
      散异矩阵 
  
    
      
        S 
        = 
        
          X 
          
            T 
           
         
        X 
        = 
        
          ∑ 
          
            i 
            = 
            1 
           
          
            n 
           
         
        
          X 
          
            ( 
            i 
            ) 
           
         
        
          X 
          
            ( 
            i 
            ) 
           
          
            T 
           
         
        , 
         
     
    {\displaystyle S=X^{T}X=\sum _{i=1}^{n}X_{(i)}X_{(i)}^{T},\,\!} 
      的几率分布 。
  
    
      
        
          S 
         
       
     
    {\displaystyle \mathbf {S} } 
      
  
    
      
        
          S 
         
        ∼ 
        
          W 
          
            p 
           
         
        ( 
        
          V 
         
        , 
        n 
        ) 
        . 
       
     
    {\displaystyle \mathbf {S} \sim W_{p}(\mathbf {V} ,n).} 
      其中正整数 
  
    
      
        n 
       
     
    {\displaystyle n} 
      自由度 。有时亦记号为
  
    
      
        W 
        ( 
        
          V 
         
        , 
        p 
        , 
        n 
        ) 
       
     
    {\displaystyle W(\mathbf {V} ,p,n)} 
      
  
    
      
        p 
        = 
        1 
       
     
    {\displaystyle p=1} 
      
  
    
      
        
          V 
         
        = 
        1 
       
     
    {\displaystyle \mathbf {V} =1} 
      
  
    
      
        n 
       
     
    {\displaystyle n} 
      卡方分布 。
 常见应用 
威沙特分布常用于多变量的概似比检定 ,亦用于随机矩阵 的频谱理论中。
 几率密度函数 
威沙特分布具有下述的几率密度函数 :
令'
  
    
      
        
          W 
         
       
     
    {\displaystyle \mathbf {W} } 
      
  
    
      
        p 
        × 
        p 
       
     
    {\displaystyle p\times p} 
      
  
    
      
        
          V 
         
       
     
    {\displaystyle \mathbf {V} } 
      
  
    
      
        p 
        × 
        p 
       
     
    {\displaystyle p\times p} 
      
如此,若
  
    
      
        n 
        > 
        p 
       
     
    {\displaystyle n>p} 
      
  
    
      
        
          W 
         
       
     
    {\displaystyle \mathbf {W} } 
      n 的威沙特分布且有几率度函数
  
    
      
        
          f 
          
            W 
           
         
       
     
    {\displaystyle f_{W}} 
      
  
    
      
        
          f 
          
            
              W 
             
           
         
        ( 
        w 
        ) 
        = 
        
          
            
              
                
                  | 
                  w 
                  | 
                 
                
                  ( 
                  n 
                  − 
                  p 
                  − 
                  1 
                  ) 
                  
                    / 
                   
                  2 
                 
               
              exp 
               
              
                [ 
                
                  − 
                  
                    
                      t 
                      r 
                      a 
                      c 
                      e 
                     
                   
                  ( 
                  
                    
                      
                        V 
                       
                     
                    
                      − 
                      1 
                     
                   
                  w 
                  
                    / 
                   
                  2 
                  ) 
                 
                ] 
               
             
            
              
                2 
                
                  n 
                  p 
                  
                    / 
                   
                  2 
                 
               
              
                
                  | 
                  
                    
                      V 
                     
                   
                  | 
                 
                
                  n 
                  
                    / 
                   
                  2 
                 
               
              
                Γ 
                
                  p 
                 
               
              ( 
              n 
              
                / 
               
              2 
              ) 
             
           
         
       
     
    {\displaystyle f_{\mathbf {W} }(w)={\frac {\left|w\right|^{(n-p-1)/2}\exp \left[-{\rm {trace}}({\mathbf {V} }^{-1}w/2)\right]}{2^{np/2}\left|{\mathbf {V} }\right|^{n/2}\Gamma _{p}(n/2)}}} 
      其中
  
    
      
        
          Γ 
          
            p 
           
         
        ( 
        ⋅ 
        ) 
       
     
    {\displaystyle \Gamma _{p}(\cdot )} 
      多变量Gamma分布 ,其定义为
  
    
      
        
          Γ 
          
            p 
           
         
        ( 
        n 
        
          / 
         
        2 
        ) 
        = 
        
          π 
          
            p 
            ( 
            p 
            − 
            1 
            ) 
            
              / 
             
            4 
           
         
        
          Π 
          
            j 
            = 
            1 
           
          
            p 
           
         
        Γ 
        
          [ 
          
            ( 
            n 
            + 
            1 
            − 
            j 
            ) 
            
              / 
             
            2 
           
          ] 
         
        . 
       
     
    {\displaystyle \Gamma _{p}(n/2)=\pi ^{p(p-1)/4}\Pi _{j=1}^{p}\Gamma \left[(n+1-j)/2\right].} 
      上述定义可推广至任一实数
  
    
      
        n 
        > 
        p 
        − 
        1 
       
     
    {\displaystyle n>p-1} 
      [2] 
 特征函数 
威沙特分布的特征函数 为
  
    
      
        Θ 
        ↦ 
        
          
            | 
            
              
                
                  I 
                 
               
              − 
              2 
              i 
              
                
                  Θ 
                 
               
              
                
                  V 
                 
               
             
            | 
           
          
            − 
            n 
            
              / 
             
            2 
           
         
        . 
       
     
    {\displaystyle \Theta \mapsto \left|{\mathbf {I} }-2i\,{\mathbf {\Theta } }{\mathbf {V} }\right|^{-n/2}.} 
      也就是说
  
    
      
        Θ 
        ↦ 
        
          
            E 
           
         
        
          { 
          
            
              e 
              x 
              p 
             
            
              [ 
              
                i 
                ⋅ 
                
                  t 
                  r 
                  a 
                  c 
                  e 
                 
                ( 
                
                  
                    W 
                   
                 
                
                  
                    Θ 
                   
                 
                ) 
               
              ] 
             
           
          } 
         
        = 
        
          
            | 
            
              
                
                  I 
                 
               
              − 
              2 
              i 
              
                
                  Θ 
                 
               
              
                
                  V 
                 
               
             
            | 
           
          
            − 
            n 
            
              / 
             
            2 
           
         
       
     
    {\displaystyle \Theta \mapsto {\mathcal {E}}\left\{\mathrm {exp} \left[i\cdot \mathrm {trace} ({\mathbf {W} }{\mathbf {\Theta } })\right]\right\}=\left|{\mathbf {I} }-2i{\mathbf {\Theta } }{\mathbf {V} }\right|^{-n/2}} 
      其中
  
    
      
        
          
            E 
           
         
        ( 
        ⋅ 
        ) 
       
     
    {\displaystyle {\mathcal {E}}(\cdot )} 
      
(这里的
  
    
      
        Θ 
       
     
    {\displaystyle \Theta } 
      
  
    
      
        
          
            I 
           
         
       
     
    {\displaystyle {\mathbf {I} }} 
      
  
    
      
        
          
            V 
           
         
       
     
    {\displaystyle {\mathbf {V} }} 
      
  
    
      
        
          
            I 
           
         
       
     
    {\displaystyle {\mathbf {I} }} 
      单位矩阵 ,而
  
    
      
        i 
       
     
    {\displaystyle i} 
      平方根 ).[3] 
 理论架构 
若
  
    
      
        
          
            
              W 
             
           
         
       
     
    {\displaystyle \scriptstyle {\mathbf {W} }} 
      m ,共变异矩阵为
  
    
      
        
          
            
              V 
             
           
         
       
     
    {\displaystyle \scriptstyle {\mathbf {V} }} 
      
  
    
      
        
          
            
              W 
             
           
          ∼ 
          
            
              
                W 
               
             
            
              p 
             
           
          ( 
          
            
              V 
             
           
          , 
          m 
          ) 
         
       
     
    {\displaystyle \scriptstyle {\mathbf {W} }\sim {\mathbf {W} }_{p}({\mathbf {V} },m)} 
      
  
    
      
        
          
            
              C 
             
           
         
       
     
    {\displaystyle \scriptstyle {\mathbf {C} }} 
      
  
    
      
        q 
        × 
        p 
       
     
    {\displaystyle q\times p} 
      q 秩矩阵,则[4] 
  
    
      
        
          
            C 
           
         
        
          
            W 
           
         
        
          
            
              C 
             
            ′ 
           
         
        ∼ 
        
          
            
              W 
             
           
          
            q 
           
         
        
          ( 
          
            
              
                C 
               
             
            
              
                V 
               
             
            
              
                
                  C 
                 
                ′ 
               
             
            , 
            m 
           
          ) 
         
        . 
       
     
    {\displaystyle {\mathbf {C} }{\mathbf {W} }{\mathbf {C} '}\sim {\mathbf {W} }_{q}\left({\mathbf {C} }{\mathbf {V} }{\mathbf {C} '},m\right).} 
      推论1 若
  
    
      
        
          
            z 
           
         
       
     
    {\displaystyle {\mathbf {z} }} 
      
  
    
      
        p 
        × 
        1 
       
     
    {\displaystyle p\times 1} 
      [4] 
  
    
      
        
          
            
              z 
             
            ′ 
           
         
        
          
            W 
           
         
        
          
            z 
           
         
        ∼ 
        
          σ 
          
            z 
           
          
            2 
           
         
        
          χ 
          
            m 
           
          
            2 
           
         
       
     
    {\displaystyle {\mathbf {z} '}{\mathbf {W} }{\mathbf {z} }\sim \sigma _{z}^{2}\chi _{m}^{2}} 
      
则在此情形下,
  
    
      
        
          χ 
          
            m 
           
          
            2 
           
         
       
     
    {\displaystyle \chi _{m}^{2}} 
      卡方分布 且
  
    
      
        
          σ 
          
            z 
           
          
            2 
           
         
        = 
        
          
            
              z 
             
            ′ 
           
         
        
          
            V 
           
         
        
          
            z 
           
         
       
     
    {\displaystyle \sigma _{z}^{2}={\mathbf {z} '}{\mathbf {V} }{\mathbf {z} }} 
      
  
    
      
        
          
            V 
           
         
       
     
    {\displaystyle {\mathbf {V} }} 
      
  
    
      
        
          σ 
          
            z 
           
          
            2 
           
         
       
     
    {\displaystyle \sigma _{z}^{2}} 
      
推论2 在
  
    
      
        
          
            
              z 
             
            ′ 
           
         
        = 
        ( 
        0 
        , 
        … 
        , 
        0 
        , 
        1 
        , 
        0 
        , 
        … 
        , 
        0 
        ) 
       
     
    {\displaystyle {\mathbf {z} '}=(0,\ldots ,0,1,0,\ldots ,0)} 
      
  
    
      
        
          w 
          
            j 
            j 
           
         
        ∼ 
        
          σ 
          
            j 
            j 
           
         
        
          χ 
          
            m 
           
          
            2 
           
         
       
     
    {\displaystyle w_{jj}\sim \sigma _{jj}\chi _{m}^{2}} 
      为矩阵的每一个对对角元素的边际分布。
统计学家George Seber 某某多变量 分布此一遣词用于所有元素的边际分布皆相同的情形。[5] 
 多变量正态分布的估计 
由于威沙特分布可视为一多变量正态分布其共变异矩阵 的最大概似估计量 (MLE)的的分布,其衍自MLE的计算可为令人惊喜地简约而优雅。[6] 频谱理论 ,可将一标量视为一
  
    
      
        1 
        × 
        1 
       
     
    {\displaystyle 1\times 1} 
      共变异矩阵的估计 。
 分布抽样 
以下的算法取材自 Smith & Hocking (1972)。[7] n 及共变异矩阵为
  
    
      
        
          V 
         
       
     
    {\displaystyle \mathbf {V} } 
      
  
    
      
        p 
        × 
        p 
       
     
    {\displaystyle p\times p} 
      
  
    
      
        n 
        ≥ 
        p 
       
     
    {\displaystyle n\geq p} 
      
生成一随机
  
    
      
        p 
        × 
        p 
       
     
    {\displaystyle p\times p} 
      三角矩阵  
  
    
      
        
          
            A 
           
         
       
     
    {\displaystyle {\textbf {A}}} 
      
  
    
      
        
          a 
          
            i 
            i 
           
         
        = 
        ( 
        
          χ 
          
            n 
            − 
            i 
            + 
            1 
           
          
            2 
           
         
        
          ) 
          
            1 
            
              / 
             
            2 
           
         
       
     
    {\displaystyle a_{ii}=(\chi _{n-i+1}^{2})^{1/2}} 
      
  
    
      
        
          a 
          
            i 
            i 
           
         
       
     
    {\displaystyle a_{ii}} 
      
  
    
      
        
          χ 
          
            n 
            − 
            i 
            + 
            1 
           
          
            2 
           
         
       
     
    {\displaystyle \chi _{n-i+1}^{2}} 
      
  
    
      
        
          a 
          
            i 
            j 
           
         
       
     
    {\displaystyle a_{ij}} 
      
  
    
      
        j 
        < 
        i 
       
     
    {\displaystyle j<i} 
      
  
    
      
        
          N 
          
            1 
           
         
        ( 
        0 
        , 
        1 
        ) 
       
     
    {\displaystyle N_{1}(0,1)} 
      正态分布 的随机样本。[8]  
计算
  
    
      
        
          
            V 
           
         
        = 
        
          
            L 
           
         
        
          
            
              L 
             
           
          
            T 
           
         
       
     
    {\displaystyle {\textbf {V}}={\textbf {L}}{\textbf {L}}^{T}} 
      Cholesky分解 。 
计算
  
    
      
        
          
            X 
           
         
        = 
        
          
            L 
           
         
        
          
            A 
           
         
        
          
            
              A 
             
           
          
            T 
           
         
        
          
            
              L 
             
           
          
            T 
           
         
       
     
    {\displaystyle {\textbf {X}}={\textbf {L}}{\textbf {A}}{\textbf {A}}^{T}{\textbf {L}}^{T}} 
      
  
    
      
        
          
            X 
           
         
       
     
    {\displaystyle {\textbf {X}}} 
      
  
    
      
        
          W 
          
            p 
           
         
        ( 
        
          
            V 
           
         
        , 
        n 
        ) 
       
     
    {\displaystyle W_{p}({\textbf {V}},n)} 
       若
  
    
      
        
          
            V 
           
         
        = 
        
          
            I 
           
         
       
     
    {\displaystyle {\textbf {V}}={\textbf {I}}} 
      
  
    
      
        
          
            V 
           
         
        = 
        
          
            I 
           
         
        
          
            
              I 
             
           
          
            T 
           
         
       
     
    {\displaystyle {\textbf {V}}={\textbf {I}}{\textbf {I}}^{T}} 
      
  
    
      
        
          
            X 
           
         
        = 
        
          
            A 
           
         
        
          
            
              A 
             
           
          
            T 
           
         
       
     
    {\displaystyle {\textbf {X}}={\textbf {A}}{\textbf {A}}^{T}} 
      
 参考条目 
共变异矩阵的估计 Hotelling的T平方分布 逆威沙特分布  参考资料 
^ Wishart, J.  The generalised product moment distribution in samples from a normal multivariate population. Biometrika . 1928, 20A  (1–2): 32–52. JFM 54.0565.02 JSTOR 2331939 doi:10.1093/biomet/20A.1-2.32   ^ Uhlig, H. On Singular Wishart and Singular Multivariate Beta Distributions . The Annals of Statistics. 1994, 22 : 395–405. doi:10.1214/aos/1176325375    ^ Anderson, T. W.  An Introduction to Multivariate Statistical Analysis  3rd. Hoboken, N. J.: Wiley Interscience . 2003: 259 . ISBN  0-471-36091-0  ^ 4.0 4.1   Rao, C. R. Linear Statistical Inference and its Applications. Wiley. 1965: 535.   ^ Seber, George A. F. Multivariate Observations. Wiley . 2004. ISBN  978-0471691211   ^ Chatfield, C.; Collins, A. J. Introduction to Multivariate Analysis . London: Chapman and Hall. 1980: 103 –108. ISBN  0-412-16030-7   ^ Smith, W. B.; Hocking, R. R. Algorithm AS 53: Wishart Variate Generator. Journal of the Royal Statistical Society, Series C . 1972, 21  (3): 341–345. JSTOR 2346290    ^ Anderson, T. W.  An Introduction to Multivariate Statistical Analysis  3rd. Hoboken, N. J.: Wiley Interscience . 2003: 257 . ISBN  0-471-36091-0   
Gelman, Andrew. Bayesian Data Analysis  2nd. Boca Raton, Fla.: Chapman & Hall. 2003: 582  [3 June  2015] . ISBN  158488388X存档 于2021-02-17).   Zanella, A.; Chiani, M.; Win, M.Z. On the marginal distribution of the eigenvalues of wishart matrices. IEEE Transactions on Communications. April 2009, 57  (4): 1050–1060. doi:10.1109/TCOMM.2009.04.070143    Bishop, C. M. Pattern Recognition and Machine Learning . Springer. 2006: 693 .   Pearson, Karl ; Jeffery, G. B. ; Elderton, Ethel M.  On the Distribution of the First Product Moment-Coefficient, in Samples Drawn from an Indefinitely Large Normal Population . Biometrika (Biometrika Trust). December 1929, 21 : 164–201. JSTOR 2332556 doi:10.2307/2332556   Craig, Cecil C. On the Frequency Function of xy . Ann. Math. Statist. 1936, 7 : 1–15  [2016-05-02 ] . doi:10.1214/aoms/1177732541 存档 于2020-06-07).   Peddada and Richards, Shyamal Das; Richards, Donald St. P. Proof of a Conjecture of M. L. Eaton on the Characteristic Function of the Wishart Distribution, . Annals of Probability . 1991, 19  (2): 868–874. doi:10.1214/aop/1176990455    Gindikin, S.G. Invariant generalized functions in homogeneous domains,. Funct. Anal. Appl.  1975, 9  (1): 50–52. doi:10.1007/BF01078179    Dwyer, Paul S. Some Applications of Matrix Derivatives in Multivariate Analysis . J. Amer. Statist. Assoc.  1967, 62  (318): 607–625. JSTOR 2283988