1;# 2;# lr.pl,v 3.1 1993/07/06 01:09:08 jbj Exp 3;# 4;# 5;# Linear Regression Package for perl 6;# to be 'required' from perl 7;# 8;# Copyright (c) 1992 9;# Frank Kardel, Rainer Pruy 10;# Friedrich-Alexander Universitaet Erlangen-Nuernberg 11;# 12;# Copyright (c) 1997 by 13;# Ulrich Windl <Ulrich.Windl@rz.uni-regensburg.de> 14;# (Converted to a PERL 5.004 package) 15;# 16;############################################################# 17 18package lr; 19 20## 21## y = A + Bx 22## 23## B = (n * Sum(xy) - Sum(x) * Sum(y)) / (n * Sum(x^2) - Sum(x)^2) 24## 25## A = (Sum(y) - B * Sum(x)) / n 26## 27 28## 29## interface 30## 31;# init(tag); initialize data set for tag 32;# sample(x, y, tag); enter sample 33;# Y(x, tag); compute y for given x 34;# X(y, tag); compute x for given y 35;# r(tag); regression coefficient 36;# cov(tag); covariance 37;# A(tag); 38;# B(tag); 39;# sigma(tag); standard deviation 40;# mean(tag); 41######################### 42 43sub init 44{ 45 my $self = shift; 46 47 $self->{n} = 0; 48 $self->{sx} = 0.0; 49 $self->{sx2} = 0.0; 50 $self->{sxy} = 0.0; 51 $self->{sy} = 0.0; 52 $self->{sy2} = 0.0; 53} 54 55sub sample($$) 56{ 57 my $self = shift; 58 my($_x, $_y) = @_; 59 60 ++($self->{n}); 61 $self->{sx} += $_x; 62 $self->{sy} += $_y; 63 $self->{sxy} += $_x * $_y; 64 $self->{sx2} += $_x**2; 65 $self->{sy2} += $_y**2; 66} 67 68sub B() 69{ 70 my $self = shift; 71 72 return 1 unless ($self->{n} * $self->{sx2} - $self->{sx}**2); 73 return ($self->{n} * $self->{sxy} - $self->{sx} * $self->{sy}) 74 / ($self->{n} * $self->{sx2} - $self->{sx}**2); 75} 76 77sub A() 78{ 79 my $self = shift; 80 81 return ($self->{sy} - B() * $self->{sx}) / $self->{n}; 82} 83 84sub Y() 85{ 86 my $self = shift; 87 88 return A() + B() * $_[$[]; 89} 90 91sub X() 92{ 93 my $self = shift; 94 95 return ($_[$[] - A()) / B(); 96} 97 98sub r() 99{ 100 my $self = shift; 101 102 my $s = ($self->{n} * $self->{sx2} - $self->{sx}**2) 103 * ($self->{n} * $self->{sy2} - $self->{sy}**2); 104 105 return 1 unless $s; 106 107 return ($self->{n} * $self->{sxy} - $self->{sx} * $self->{sy}) / sqrt($s); 108} 109 110sub cov() 111{ 112 my $self = shift; 113 114 return ($self->{sxy} - $self->{sx} * $self->{sy} / $self->{n}) 115 / ($self->{n} - 1); 116} 117 118sub sigma() 119{ 120 my $self = shift; 121 122 return 0 if $self->{n} <= 1; 123 return sqrt(($self->{sy2} - ($self->{sy} * $self->{sy}) / $self->{n}) 124 / ($self->{n})); 125} 126 127sub mean() 128{ 129 my $self = shift; 130 131 return 0 if $self->{n} <= 0; 132 return $self->{sy} / $self->{n}; 133} 134 135sub new 136{ 137 my $class = shift; 138 my $self = { 139 (n => undef, 140 sx => undef, 141 sx2 => undef, 142 sxy => undef, 143 sy => undef, 144 sy2 => undef) 145 }; 146 bless $self, $class; 147 init($self); 148 return $self; 149} 150 1511; 152