Class: MDS::StdlibInterface

Inherits:
MatrixInterface show all
Defined in:
lib/mds/interfaces/stdlib_interface.rb

Overview

Matrix interface for Ruby’s standard library Matrix class.

To succesfully use this interface ‘extendmatrix’ is required. For more information and installation instructions see github.com/clbustos/extendmatrix

The algorithm/implementation of the eigen-decomposition is sub-optimal. It is therefore suitable only for small-scale problems. Timings are illustrated in the benchmark section of the README.

The reason this interface is still included in RMDS is the fact that it only depends on Ruby code and no native extensions.

Compatible with ‘extendmatrix >= 0.3.1’

Class Method Summary collapse

Methods inherited from MatrixInterface

columns, create_block, create_diagonal, create_identity, create_random, create_rows, diagonals, inherited, minor, rows, #to_s, trace

Class Method Details

.add(m, n) ⇒ Object

Componentwise addition of two matrices.



114
115
116
 
# File 'lib/mds/interfaces/stdlib_interface.rb', line 114

def StdlibInterface.add(m, n)
  m + n
end

.create(n, m, s) ⇒ Object

Create a new matrix with equal elements.



44
45
46
 
# File 'lib/mds/interfaces/stdlib_interface.rb', line 44

def StdlibInterface.create(n, m, s)
  ::Matrix.build(n, m, s)
end

.ed(m) ⇒ Object

Compute the eigen-decomposition of a real symmetric matrix.

The Ruby version uses Jacobi iterations to calculate the eigen-decomposition of a matrix. Although comfortable as all third-party dependencies can be installed via gem, the method is not suited for matrices bigger than 10x10.



139
140
141
142
143
144
145
146
147
148
149
150
151
152
 
# File 'lib/mds/interfaces/stdlib_interface.rb', line 139

def StdlibInterface.ed(m)
  eigen_values = m.cJacobiA
  eigen_vectors = m.cJacobiV

  ranks = (0..(m.row_size-1)).sort{|i,j| eigen_values[j,j] <=> eigen_values[i,i]}
  
  s_eigen_values = []
  s_eigen_vectors = []
  ranks.each do |r|
    s_eigen_values << eigen_values[r,r]
    s_eigen_vectors << eigen_vectors.column(r)
  end    
  [::Matrix.diagonal(*s_eigen_values), ::Matrix.columns(s_eigen_vectors)]
end

.get(m, i, j) ⇒ Object

Get matrix element.



83
84
85
 
# File 'lib/mds/interfaces/stdlib_interface.rb', line 83

def StdlibInterface.get(m, i, j)
  m[i,j]
end

.ncols(m) ⇒ Object

Return the number of matrix columns



64
65
66
 
# File 'lib/mds/interfaces/stdlib_interface.rb', line 64

def StdlibInterface.ncols(m)
  m.column_size
end

.nrows(m) ⇒ Object

Return the number of matrix rows



54
55
56
 
# File 'lib/mds/interfaces/stdlib_interface.rb', line 54

def StdlibInterface.nrows(m)
  m.row_size
end

.prod(m, n) ⇒ Object

Calculate the product of two matrices or the product of a matrix and a scalar.



94
95
96
 
# File 'lib/mds/interfaces/stdlib_interface.rb', line 94

def StdlibInterface.prod(m, n)
  m * n
end

.set(m, i, j, s) ⇒ Object

Set matrix element.



73
74
75
 
# File 'lib/mds/interfaces/stdlib_interface.rb', line 73

def StdlibInterface.set(m, i, j, s)
  m[i,j] = s
end

.sub(m, n) ⇒ Object

Componentwise subtraction of two matrices.



124
125
126
 
# File 'lib/mds/interfaces/stdlib_interface.rb', line 124

def StdlibInterface.sub(m, n)
  m - n
end

.t(m) ⇒ Object

Transpose a matrix.



104
105
106
 
# File 'lib/mds/interfaces/stdlib_interface.rb', line 104

def StdlibInterface.t(m)
  m.t
end