wedesoft / linalg Goto Github PK
View Code? Open in Web Editor NEWThis project forked from quix/linalg
Ruby Linear Algebra Library
Home Page: http://linalg.rubyforge.org
License: Other
This project forked from quix/linalg
Ruby Linear Algebra Library
Home Page: http://linalg.rubyforge.org
License: Other
= Linalg - Ruby Linear Algebra Library A Fortran-based linear algebra package. === Features Major features: * Cholesky decomposition * LU decomposition * QR decomposition * Schur decomposition * Singular value decomposition * Eigenvalues and eigenvectors of a general matrix * Minimization by least squares * Linear equation solving * Stand-alone LAPACK bindings: call any LAPACK routine from directly from ruby. Minor features: * Convenient iterators * Condition numbers and condition number estimates * Nullspace, rank, nullity * Inverse * Pseudo-inverse * Determinant * 2-norm, 1-norm, infinity-norm, Frobenius norm === Getting Started Everything you need to know is in: * Linalg::DMatrix * Linalg::Iterators and this README. === Tutorial $ irb irb(main):000:0> require 'linalg' => true irb(main):000:0> include Linalg => Object ==== Construction irb(main):000:0> DMatrix[[1,2,3], [4,5,6]] => 1.000000 2.000000 3.000000 4.000000 5.000000 6.000000 irb(main):000:0> DMatrix.rows [[1,2,3], [4,5,6]] => 1.000000 2.000000 3.000000 4.000000 5.000000 6.000000 irb(main):000:0> DMatrix.columns [[1,2,3], [4,5,6]] => 1.000000 4.000000 2.000000 5.000000 3.000000 6.000000 irb(main):000:0> a = DMatrix.new(3, 3) { |i, j| 10*i + j } => 0.000000 1.000000 2.000000 10.000000 11.000000 12.000000 20.000000 21.000000 22.000000 irb(main):000:0> DMatrix.new(3, 3, 99) => 99.000000 99.000000 99.000000 99.000000 99.000000 99.000000 99.000000 99.000000 99.000000 irb(main):000:0> DMatrix.diagonal [3,4,5] => 3.000000 0.000000 0.000000 0.000000 4.000000 0.000000 0.000000 0.000000 5.000000 irb(main):000:0> DMatrix.diagonal(4) { |i| i*i } => 0.000000 0.000000 0.000000 0.000000 0.000000 1.000000 0.000000 0.000000 0.000000 0.000000 4.000000 0.000000 0.000000 0.000000 0.000000 9.000000 irb(main):000:0> DMatrix.diagonal(4, 99) => 99.000000 0.000000 0.000000 0.000000 0.000000 99.000000 0.000000 0.000000 0.000000 0.000000 99.000000 0.000000 0.000000 0.000000 0.000000 99.000000 ==== Indexing Indexing is in (row, column) order. This is the convention for Mathematics and Fortran. It is opposite from C convention. irb(main):000:0> a => 0.000000 1.000000 2.000000 10.000000 11.000000 12.000000 20.000000 21.000000 22.000000 irb(main):000:0> a[1,0] => 10.0 irb(main):000:0> a[2,0] => 20.0 irb(main):000:0> a[0,1] => 1.0 irb(main):000:0> a[0,2] => 2.0 Index boundaries are strongly enforced irb(main):000:0> a[-1,0] IndexError: out of range from (irb):27:in `[]' from (irb):27 ==== Enumerables There are several abstract <tt>Enumerable</tt>s which you may obtain from a matrix: columns, rows, elements, and diagonal elements. irb(main):000:0> a => 0.000000 1.000000 2.000000 10.000000 11.000000 12.000000 20.000000 21.000000 22.000000 irb(main):000:0> a.columns.class => Linalg::Iterators::ColumnEnum irb(main):000:0> cols = a.columns.map { |x| x } => [ 0.000000 10.000000 20.000000 , 1.000000 11.000000 21.000000 , 2.000000 12.000000 22.000000 ] irb(main):000:0> rows = a.rows.map { |x| x } => [ 0.000000 1.000000 2.000000 , 10.000000 11.000000 12.000000 , 20.000000 21.000000 22.000000 ] irb(main):000:0> a.elems.map { |x| x } => [0.0, 10.0, 20.0, 1.0, 11.0, 21.0, 2.0, 12.0, 22.0] irb(main):003:0> a.elems.find_all { |x| x > 10 } => [20.0, 11.0, 21.0, 12.0, 22.0] irb(main):008:0> a.diags.map { |x| x } => [0.0, 11.0, 22.0] Another method of constructing a matrix is to join rows or columns, irb(main):000:0> DMatrix.join_columns [cols[0], cols[2]] => 0.000000 2.000000 10.000000 12.000000 20.000000 22.000000 irb(main):000:0> DMatrix.join_rows [rows[0], rows[2]] => 0.000000 1.000000 2.000000 20.000000 21.000000 22.000000 ==== Enumerable-like Iterators with Index Pairs A matrix itself is not +Enumerable+, but a select number of +Enumerable+-like methods are provided. irb(main):000:0> a => 0.000000 1.000000 2.000000 10.000000 11.000000 12.000000 20.000000 21.000000 22.000000 irb(main):000:0> a.each_with_index { |e, i, j| puts "row #{i} column #{j} : #{e}" } ; nil row 0 column 0 : 0.0 row 1 column 0 : 10.0 row 2 column 0 : 20.0 row 0 column 1 : 1.0 row 1 column 1 : 11.0 row 2 column 1 : 21.0 row 0 column 2 : 2.0 row 1 column 2 : 12.0 row 2 column 2 : 22.0 => nil irb(main):000:0> a.map_with_index { |e, i, j| e*i*j } => 0.000000 0.000000 0.000000 0.000000 11.000000 24.000000 0.000000 42.000000 88.000000 irb(main):000:0> a.each_upper_with_index { |e, i, j| puts "a[#{i}, #{j}] : #{e}" } ; nil a[0, 1] : 1.0 a[0, 2] : 2.0 a[1, 2] : 12.0 => nil irb(main):000:0> a.each_lower_with_index { |e, i, j| puts "a[#{i}, #{j}] : #{e}" } ; nil a[1, 0] : 10.0 a[2, 0] : 20.0 a[2, 1] : 21.0 => nil ==== Epsilon Comparison For good and bad, a default epsilon of <tt>1e-8</tt> is provided for comparison, nullspace identification, and symmetric testing. You can change +default_epsilon+ class-wide or on a per-object basis, or simply pass an explicit epsilon to any of these methods. irb(main):000:0> a = DMatrix.rand(3, 3) => 0.824730 0.305527 0.044433 -0.582865 -0.351364 -0.752941 0.103417 -0.254290 0.216312 irb(main):000:0> b = a.map { |e| e + 0.000001 } => 0.824731 0.305528 0.044434 -0.582864 -0.351363 -0.752940 0.103418 -0.254289 0.216313 irb(main):000:0> a.within(1e-4, b) => true irb(main):000:0> a.class.default_epsilon => 1.0e-08 irb(main):000:0> a =~ b => false irb(main):000:0> a.singleton_class.default_epsilon => nil irb(main):000:0> a.singleton_class.default_epsilon = 0.0001 => 0.0001 irb(main):000:0> a =~ b => true irb(main):000:0> b =~ a => false <tt>singleton_class.epsilon</tt> has first preference over <tt>class.epsilon</tt>. ==== Singular Value Decomposition irb(main):000:0> a = DMatrix.rand(4, 7) ; irb(main):000:0* u, s, vt = a.singular_value_decomposition => [ -0.747003 0.304315 -0.144972 -0.573029 -0.435034 -0.814506 0.381951 0.037926 0.207010 -0.490811 -0.777727 -0.333753 -0.458125 0.055467 -0.477741 0.747535 , 2.186983 0.000000 0.000000 0.000000 0.000000 0.000000 0.000000 0.000000 1.719562 0.000000 0.000000 0.000000 0.000000 0.000000 0.000000 0.000000 1.474243 0.000000 0.000000 0.000000 0.000000 0.000000 0.000000 0.000000 0.676138 0.000000 0.000000 0.000000 , 0.276463 -0.345917 0.573929 -0.416910 -0.139648 0.526564 0.062694 -0.838456 -0.430716 0.199266 0.170144 0.087954 0.077829 0.170372 -0.022382 -0.403705 -0.054483 -0.077530 -0.042506 -0.160232 -0.894461 -0.079171 0.476099 0.396614 0.253131 0.573392 0.315335 -0.342737 0.034285 -0.212892 -0.265934 -0.537325 0.749204 -0.111692 0.142407 -0.377150 0.344083 -0.443538 -0.422834 -0.235075 0.530130 -0.165989 -0.265280 0.376113 0.450755 -0.509553 -0.160031 -0.545940 -0.041002 ] irb(main):000:0> u*s*vt => -0.854950 0.241548 -0.975366 0.688628 0.061092 -0.907441 0.310691 0.896671 0.717254 -0.845643 0.121186 0.000445 -0.692125 -0.810721 0.876331 0.562343 0.064623 -0.300574 -0.218113 0.285261 0.987489 -0.381214 0.830467 -0.317184 0.616481 0.468055 -0.247913 0.410178 irb(main):000:0> a => -0.854950 0.241548 -0.975366 0.688628 0.061092 -0.907441 0.310691 0.896671 0.717254 -0.845643 0.121186 0.000445 -0.692125 -0.810721 0.876331 0.562343 0.064623 -0.300574 -0.218113 0.285261 0.987489 -0.381214 0.830467 -0.317184 0.616481 0.468055 -0.247913 0.410178 irb(main):000:0> u*u.t => 1.000000 0.000000 -0.000000 0.000000 0.000000 1.000000 0.000000 0.000000 -0.000000 0.000000 1.000000 0.000000 0.000000 0.000000 0.000000 1.000000 irb(main):000:0> vt.t*vt => 1.000000 0.000000 0.000000 0.000000 0.000000 0.000000 -0.000000 0.000000 1.000000 -0.000000 -0.000000 0.000000 -0.000000 0.000000 0.000000 -0.000000 1.000000 -0.000000 -0.000000 0.000000 0.000000 0.000000 -0.000000 -0.000000 1.000000 -0.000000 -0.000000 0.000000 0.000000 0.000000 -0.000000 -0.000000 1.000000 -0.000000 0.000000 0.000000 -0.000000 0.000000 -0.000000 -0.000000 1.000000 -0.000000 -0.000000 0.000000 0.000000 0.000000 0.000000 -0.000000 1.000000 ==== Eigenvectors and Eigenvalues irb(main):000:0> a = DMatrix.rand(5, 5) => -0.319566 0.633985 0.335298 -0.150403 0.758559 -0.633389 0.444269 0.375873 0.521107 0.247966 0.757654 0.504831 0.160970 -0.241885 -0.949746 -0.174517 0.351239 -0.600079 -0.533921 0.851118 -0.736717 0.006612 -0.941311 -0.417801 0.555841 irb(main):000:0> eigs, re, im = a.eigensystem => [ -0.232169 -0.540550 -0.215201 -0.216959 0.036677 -0.449048 -0.031420 -0.114366 -0.228818 0.062841 0.682632 0.283575 0.019432 0.540907 0.073854 0.364843 -0.597905 0.000000 -0.592036 0.000000 0.381257 -0.207285 -0.407649 -0.490737 -0.076896 , -1.088525 -0.093563 -0.093563 0.791621 0.791621 , 0.000000 0.604163 -0.604163 0.158934 -0.158934 ] irb(main):000:0> a*eigs.column(0) => 0.252722 0.488799 -0.743062 -0.397141 -0.415008 irb(main):000:0> re[0]*eigs.column(0) => 0.252722 0.488799 -0.743062 -0.397141 -0.415008 ==== QR factorization irb(main):000:0> a = DMatrix.rand(4, 7) ; irb(main):000:0* q, r = a.qr => [ -0.593983 0.263367 -0.572195 -0.500414 -0.486715 -0.780258 0.331047 -0.211458 -0.184137 -0.328403 -0.586889 0.716803 -0.613503 0.462588 0.467507 0.437109 , -1.180464 -0.295897 -0.478813 -0.300434 0.360094 0.738799 -0.571687 0.000000 0.873800 -0.167873 0.612545 0.462280 0.097813 -0.733398 0.000000 0.000000 1.659791 -0.439355 -0.234388 -0.063566 0.149709 0.000000 0.000000 0.000000 -0.047795 0.805122 -0.158261 0.754104 ] irb(main):000:0> q*r => 0.701176 0.405888 -0.709530 0.615091 -0.360919 -0.297505 -0.316607 0.574550 -0.537772 0.913498 -0.467057 -0.783803 -0.423482 0.740588 0.217367 -0.232473 -0.830816 0.077752 0.496553 -0.244298 0.798800 0.724218 0.585743 0.992061 0.241379 0.235274 -0.506903 0.411086 irb(main):000:0> a => 0.701176 0.405888 -0.709530 0.615091 -0.360919 -0.297505 -0.316607 0.574550 -0.537772 0.913498 -0.467057 -0.783803 -0.423482 0.740588 0.217367 -0.232473 -0.830816 0.077752 0.496553 -0.244298 0.798800 0.724218 0.585743 0.992061 0.241379 0.235274 -0.506903 0.411086 irb(main):000:0> q.t*q => 1.000000 0.000000 -0.000000 -0.000000 0.000000 1.000000 0.000000 -0.000000 -0.000000 0.000000 1.000000 0.000000 -0.000000 -0.000000 0.000000 1.000000 See the Linalg::DMatrix documentation for more info. The various tests in test/ are also instructive. === Download * http://rubyforge.org/frs/?group_id=273 === Repository * http://github.com/quix/linalg === Notes There are four matrix types: +SMatrix+, +DMatrix+, +CMatrix+, and +ZMatrix+ -- single precision, double precision, single precision complex, and double precision complex, respectively. They are all available with basic functionality, however the more complex routines you see here currently lie only in +DMatrix+. If you have used +narray+, note that +linalg+ uses the mathematical definition of <em>rank</em>, which is equal to the number of columns only in the case of a nonsingular square matrix. === Details Author:: James M. Lawrence <[email protected]> Requires:: Ruby 1.8.1 or later License:: Copyright (c) 2004-2008 James M. Lawrence. Released under the MIT license. === License Copyright (c) 2004-2008 James M. Lawrence If +linalg+ begins to smoke, get away immediately. Seek shelter and cover head. Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the "Software"), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so, subject to the following conditions: The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software. THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
A declarative, efficient, and flexible JavaScript library for building user interfaces.
๐ Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
An Open Source Machine Learning Framework for Everyone
The Web framework for perfectionists with deadlines.
A PHP framework for web artisans
Bring data to life with SVG, Canvas and HTML. ๐๐๐
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
Some thing interesting about web. New door for the world.
A server is a program made to process requests and deliver data to clients.
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
Some thing interesting about visualization, use data art
Some thing interesting about game, make everyone happy.
We are working to build community through open source technology. NB: members must have two-factor auth.
Open source projects and samples from Microsoft.
Google โค๏ธ Open Source for everyone.
Alibaba Open Source for everyone
Data-Driven Documents codes.
China tencent open source team.