A preliminary report on the development of MATLAB tensor classes for fast algorithm prototyping.

Bader, Brett William ; Kolda, Tamara Gibson (Sandia National Laboratories, Livermore, CA)

Sandia National Laboratories

关键词: 99 General And Miscellaneous//Mathematics, Computing, And Information Science;    97 Mathematical Methods And Computing;    Calculus Of Tensors.;    Algorithms;    Computer Codes;   

DOI  :  10.2172/974887
RP-ID  :  SAND2004-3487
RP-ID  :  AC04-94AL85000
RP-ID  :  974887

美国|英语

来源: UNT Digital Library

【 摘 要 】

We describe three MATLAB classes for manipulating tensors in order to allow fast algorithm prototyping. A tensor is a multidimensional or N-way array. We present a tensor class for manipulating tensors which allows for tensor multiplication and 'matricization.' We have further added two classes for representing tensors in decomposed format: cp{_}tensor and tucker{_}tensor. We demonstrate the use of these classes by implementing several algorithms that have appeared in the literature.