comfac

  comfac算法是三线性模型常见的数据检测方法。它首先将三线性数据通过tucker3模型压缩，然后通过最小二乘的方法不断更新数据，直到收敛，得到估计值( The algorithm works by first compressing the data using a Tucker3 models. Subsequently the PARAFAC model is fitted to the compressed array, either initialized with DTLD (~ESPRIT) or with PARAFAC-ALS. The solution is de-compressed to the original space, and a few safe-guard PARAFAC-ALS steps are performed. Further optimization of this algorithm is possible, e.g. if the data are very large, if the profiles are very correlated or if the noise is huge. This 、 has not been pursued here, in order to have a generally applicable algorithm)