matlab 中tt_fac_to_vec函数

布满荆棘的人生 2022-05-26 14:05 103阅读 0赞

**语法：**

X = tt\_fac\_to\_vec(A)

converts a cell array of factor matrices A to a  vector by vectorizing each matrix and stacking them.

函数功能就是将因子矩阵构成的细胞数组A 通过向量化每个数组 并且叠放一起 转化成一个向量。

**例子**：

%生成随机的2\*3\*2的张量

>> a=tensor(rand(2,3,2))  
a is a tensor of size 2 x 3 x 2  
a(:,:,1) =   
    0.4908    0.2657    0.0638  
    0.7251    0.2206    0.5931  
a(:,:,2) =   
    0.7621    0.5358    0.2929

0.1999    0.6221    0.2121

%对张量进行cp分解

>> b=cp\_als(a,2)  
CP\_ALS:  
 Iter  1: f = 6.623389e-01 f-delta = 6.6e-01  
 Iter  2: f = 7.614281e-01 f-delta = 9.9e-02  
 Iter  3: f = 8.049933e-01 f-delta = 4.4e-02  
 Iter  4: f = 8.154584e-01 f-delta = 1.0e-02  
 Iter  5: f = 8.205675e-01 f-delta = 5.1e-03  
 Iter  6: f = 8.232252e-01 f-delta = 2.7e-03  
 Iter  7: f = 8.246533e-01 f-delta = 1.4e-03  
 Iter  8: f = 8.254496e-01 f-delta = 8.0e-04  
 Iter  9: f = 8.259128e-01 f-delta = 4.6e-04  
 Iter 10: f = 8.261940e-01 f-delta = 2.8e-04  
 Iter 11: f = 8.263714e-01 f-delta = 1.8e-04  
 Iter 12: f = 8.264871e-01 f-delta = 1.2e-04  
 Iter 13: f = 8.265644e-01 f-delta = 7.7e-05  
 Final f = 8.265644e-01   
b is a ktensor of size 2 x 3 x 2  
b.lambda = \[ 1.5422     0.72091 \]  
b.U\{1\} =   
    0.7211    0.0195  
    0.6928    0.9998  
b.U\{2\} =   
    0.7293    0.8234  
    0.5599   -0.1567  
    0.3932    0.5455  
b.U\{3\} =   
    0.4739    0.7184  
    0.8806   -0.6956  
%因为因子矩阵在b.U的细胞数组总存储，使用tt\_fac\_to\_vec函数对其进行向量化。  
>> c=tt\_fac\_to\_vec(b.U)   
c =  
    0.7211  
    0.6928  
    0.0195  
    0.9998  
    0.7293  
    0.5599  
    0.3932  
    0.8234  
   -0.1567  
    0.5455  
    0.4739  
    0.8806  
    0.7184  
   -0.6956