Discovery of factors in binary data triangular decomposition of matrices
نویسنده
چکیده
We present new methods of decomposition of an n ×m binary matrix I into a product A ∗ B of an n × k binary matrix A and a k×m binary matrix B. These decompositions are alternative to the usual one which is sought in Boolean factor analysis (BFA), where ∗ is a Boolean product of matrices. In the new decompositions, ∗ are the left and the right triangular products of Boolean matrices. In BFA, I is interpreted as object × attribute matrix, A and B are interpreted as object × factor and factor × attribute matrices, and the aim is to find a decomposition with the number k of factors as small as possible. The new decompositions have different semantics from the one with Boolean matrix product. The presented methods are optimal in that they provide the smallest number k possible. We provide a geometric insight showing that the factors correspond to I-beamand H-beam-shaped patterns in the input matrix. We present an approximation algorithm for computing the optimal decomposition. The algorithm is based on a greedy approximation algorithm for set covering optimization problem. We demonstrate our results by illustrative examples.
منابع مشابه
Optimal triangular decompositions of matrices with entries from residuated lattices
We describe optimal decompositions of an n m matrix I into a triangular product I 1⁄4 A / B of an n k matrix A and a k m matrix B. We assume that the matrix entries are elements of a residuated lattice, which leaves binary matrices or matrices which contain numbers from the unit interval [0,1] as special cases. The entries of I, A, and B represent grades to which objects have attributes, factor...
متن کاملDiscovery of optimal factors in binary data via a novel method of matrix decomposition
a r t i c l e i n f o a b s t r a c t Keywords: Binary matrix decomposition Factor analysis Binary data Formal concept analysis Concept lattice We present a novel method of decomposition of an n × m binary matrix I into a Boolean product A • B of an n × k binary matrix A and a k × m binary matrix B with k as small as possible. Attempts to solve this problem are known from Boolean factor analysi...
متن کاملJoint and Generalized Spectral Radius of Upper Triangular Matrices with Entries in a Unital Banach Algebra
In this paper, we discuss some properties of joint spectral {radius(jsr)} and generalized spectral radius(gsr) for a finite set of upper triangular matrices with entries in a Banach algebra and represent relation between geometric and joint/generalized spectral radius. Some of these are in scalar matrices, but some are different. For example for a bounded set of scalar matrices,$Sigma$, $r_*...
متن کاملTRIANGULAR FUZZY MATRICES
In this paper, some elementary operations on triangular fuzzynumbers (TFNs) are defined. We also define some operations on triangularfuzzy matrices (TFMs) such as trace and triangular fuzzy determinant(TFD). Using elementary operations, some important properties of TFMs arepresented. The concept of adjoints on TFM is discussed and some of theirproperties are. Some special types of TFMs (e.g. pu...
متن کاملMatrix Decomposition of the Unified Generalized Stirling Numbers and Inversion of the Generalized Factorial Matrices
In this paper, we give a matrix decomposition method used to calculate unified generalized Stirling numbers in an explicit, non-recursive mode, and some of its applications. Then, we define generalized factorial matrices which may be regarded as a generalization in the form of the Vandermonde matrices, and presents some of their properties — in particular, triangular matrix factors of the inver...
متن کامل