What is band storage of matrix? - algorithm

I was asked about various storage of matrices; in particular, about band storage and other variations. I know it is something related to storage of sparse matrices in an efficient way. But, I have no clear idea about the details.

Simply put, a matrix is banded if all the non-zero elements are close to the diagonal. More formally, consider a matrix A whose elements are aij. The matrix is said to be banded with bandwith m if aij = 0 for all i, j such that abs(i-j)>=m.
A banded matrix is a sparse matrix of a very special form. Band structure is very easy to understand and operate on. Storage is efficient, and very efficient algorithms exist for banded matrices. More so than for more general sparse matrices.
Band storage takes advantage of the structure of the matrices by only storing the elements that may be non-zero.
Read more about this here: http://en.wikipedia.org/wiki/Band_matrix

Related

Precise matrix inversion in Q

Given an invertible matrix M over the rationals Q, the inverse matrix M^(-1) is again a matrix over Q. Are their (efficient) libraries to compute the inverse precisely?
I am aware of high-performance linear algebra libraries such as BLAS/LAPACK, but these libraries are based on floating point arithmetic and are thus not suitable for computing precise (analytical) solutions.
Motivation: I want to compute the absorption probabilities of a large absorbing Markov chain using its fundamental matrix. I would like to do so precisely.
Details: By large, I mean a 1000x1000 matrix in the best case, and a several million dimensional matrix in the worst case. The further I can scale things the better. (I realize that the worst case is likely far out of reach.)
You can use the Eigen matrix library, which with little effort works on arbitrary scalar types. There is an example in the documentation how to use it with GMPs mpq_class: http://eigen.tuxfamily.org/dox/TopicCustomizing_CustomScalar.html
Of course, as #btilly noted, most of the time you should not calculate the inverse, but calculate a matrix decomposition and use that to solve equation systems. For rational numbers you can use any LU-decomposition, or if the matrix is symmetric, the LDLt decomposition. See here for a catalogue of decompositions.

What is the best way to multiply a large dense matrix with its transpose?

I have a large matrix of the order 1M x 300 (obtained after SVD decomposition of a large item matrix). So, the matrix is a dense one with float as data type. I would like to compute the similarity matrix by multiplying the dimensionally reduced matrix with its transpose.
I implemented the matrix multiplication method and that doesn't just end.
What are the ways to perform matrix multiplication between the dense matrix (~1M rows x 300 columns) with its transpose?
Will using MapReduce help in speeding up the job?
I also saw Apache Hama being efficient for large matrix computations. Will that fit my problem?
Strassen's algorithm is also used for large matrices, how do i use it?
Any other solutions/suggestion for it?

Matrix-Vector Multiplication - Sparse vs. Dense matrices

I want to implement a matrix-vector multiplication in C. My matrix is 1000 * 1000^2 and highly sparse (less than 0.01% non-zero elements). The non-zero elements are dispersed among the rows (between 0 to 126 non-zero elements per row).
I have heard that generally, using parallel processing for sparse matrix-vector multiplication is challenging and not as efficient as dense matrices because the ratio of computation to memory access is low (Here). But I cannot really understand what is the main difference between a sparse and a dense matrix with respect to parallel computation that makes sparse matrices less efficient.It seems the same problem is still around for the dense matrices (please correct me if I am wrong).
It is appreciated if let me know how dense matrices differ from sparse matrices in terms of parallel processing.
Thanks

Algorithms for Performing Large Integer Matrix Operations w/ Numerical Stability

I'm looking for a library that performs matrix operations on large sparse matrices w/o sacrificing numerical stability. Matrices will be 1000+ by 1000+ and values of the matrix will be between 0 and 1000. I will be performing the index calculus algorithm (en.wikipedia.org/wiki/Index_calculus_algorithm) so I will be generating (sparse) row vectors of the matrix serially. As I develop each row, I will need to test for linear independence. Once I fill my matrix with the desired number of linearly independent vectors, I will then need to transform the matrix into reduced row echelon form.
The problem now is that my implementation uses Gaussian elimination to determine linear independence (ensuring row echelon form once all my row vectors have been found). However, given the density and size of the matrix, this means the entries in each new row become exponentially larger over time, as the lcm of the leading entries must be found in order to perform cancellation. Finding the reduced form of the matrix further exacerbates the problem.
So my question is, is there an algorithm, or better yet an implementation, that can test linear independence and solve the reduced row echelon form while keeping the entries as small as possible? An efficient test for linear independence is especially important since in the index calculus algorithm it is performed by far the most.
Thanks in advance!
Usually if you are working with large matrices, people use LAPACK: this library contains all the basic matrix routines and support many different matrix types (sparse, ...). You can use this library to implement your algorithm, I think it will help you

Efficient algorithm for finding largest eigenpair of small general complex matrix

I am looking for an efficient algorithm to find the largest eigenpair of a small, general (non-square, non-sparse, non-symmetric), complex matrix, A, of size m x n. By small I mean m and n is typically between 4 and 64 and usually around 16, but with m not equal to n.
This problem is straight forward to solve with the general LAPACK SVD algorithms, i.e. gesvd or gesdd. However, as I am solving millions of these problems and only require the largest eigenpair, I am looking for a more efficient algorithm. Additionally, in my application the eigenvectors will generally be similar for all cases. This lead me to investigate Arnoldi iteration based methods, but I have neither found a good library nor algorithm that applies to my small general complex matrix. Is there an appropriate algorithm and/or library?
Rayleigh iteration has cubic convergence. You may want to implement also the power method and see how they compare, since you need LU or QR decomposition of your matrix.
http://en.wikipedia.org/wiki/Rayleigh_quotient_iteration
Following #rchilton's comment, you can apply this to A* A.
The idea of looking for the largest eigenpair is analogous to finding a large power of the matrix, as the lower frequency modes get damped out during the iteration. The Lanczos algorithm, is one of a few such algorithms that rely on the so-called Ritz eigenvectors during the decomposition. From Wikipedia:
The Lanczos algorithm is an iterative algorithm ... that is an adaptation of power methods to find eigenvalues and eigenvectors of a square matrix or the singular value decomposition of a rectangular matrix. It is particularly useful for finding decompositions of very large sparse matrices. In latent semantic indexing, for instance, matrices relating millions of documents to hundreds of thousands of terms must be reduced to singular-value form.
The technique works even if the system is not sparse, but if it is large and dense it has the advantage that it doesn't all have to be stored in memory at the same time.
How does it work?
The power method for finding the largest eigenvalue of a matrix A can be summarized by noting that if x_{0} is a random vector and x_{n+1}=A x_{n}, then in the large n limit, x_{n} / ||x_{n}|| approaches the normed eigenvector corresponding to the largest eigenvalue.
Non-square matrices?
Noting that your system is not a square matrix, I'm pretty sure that the SVD problem can be decomposed into separate linear algebra problems where the Lanczos algorithm would apply. A good place to ask such questions would be over at https://math.stackexchange.com/.

Resources