1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
|
/*
gblreg (ver. 5.6) -- Estimate linear regression model
Copyright (C) 2005-2018 Giulio Bottazzi
This program is free software; you can redistribute it and/or
modify it under the terms of the GNU General Public License
(version 2) as published by the Free Software Foundation;
This program is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
GNU General Public License for more details.
You should have received a copy of the GNU General Public License
along with this program; if not, write to the Free Software
Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
*/
#include "tools.h"
#include <gsl/gsl_fit.h>
#include <gsl/gsl_cdf.h>
void print_linear_verbose(FILE *file,const int o_weight, const int o_model,
const size_t size,
const double c0,const double c1,
const double cov00,const double cov01, const double cov11,
const double sumsq){
const int ndf = size-(o_model==0?2:1);
fprintf(file,"\n");
fprintf(file," ------------------------------------------------------------\n");
fprintf(stderr," number of observations = %zd\n",size);
fprintf(stderr," number of indep. variables = %d\n",ndf);
fprintf(file,"\n");
fprintf(stderr," model (C[i]: i-th column; eps: residual):\n\n");
fprintf(stderr," C[2] ~ ");
if(o_model==0) fprintf(stderr," c0 + ");
fprintf(stderr,"c1*C[1] + ");
if(o_weight)
fprintf(stderr,"eps*C[3]\n");
else
fprintf(stderr,"eps\n");
fprintf(file,"\n");
fprintf(file," ------------------------------------------------------------\n");
if(!o_weight){
fprintf(file," sum of squared residual (SSR) = %f\n",sumsq);
fprintf(file," number degrees of freedom (ndf) = %d\n",ndf);
fprintf(file," sqrt(SSR/ndf) = %f\n",sqrt(sumsq/ndf));
fprintf(file," chi-square test P(>SSR | ndf) = %e\n",
gsl_cdf_chisq_Q (sumsq,ndf));
}
else{
fprintf(file," sum of weighted squared residual (WSSR) = %f\n",sumsq);
fprintf(file," number degrees of freedom (ndf) = %d\n",ndf);
fprintf(file," sqrt(WSSR/ndf) = %f\n",sqrt(sumsq/ndf));
fprintf(file," chi-square test P(>WSSR | ndf) = %e\n",
gsl_cdf_chisq_Q (sumsq,ndf));
}
fprintf(file," ------------------------------------------------------------\n");
fprintf(file,"\n");
/* Normal approximation to chi^2 */
/* fprintf(file," chisq. test Q(ndf/2,WSSR/2) = %e\n", */
/* gsl_cdf_gaussian_Q(sumsq-(size-cov->size1),sqrt(2*(size-cov->size1)))); */
if(o_model==0){
fprintf(file," c0= %+f +/- %f (%5.1f%%) | % f % f |\n",c0,sqrt(cov00),100*sqrt(cov00)/fabs(c0),1.,cov01/sqrt(cov00*cov11));
fprintf(file," c1= %+f +/- %f (%5.1f%%) | % f % f |\n",c1,sqrt(cov11),100*sqrt(cov11)/fabs(c1),cov01/sqrt(cov00*cov11),1.);
}
else {
fprintf(file," c1= %+f +/- %f (%5.1f%%)\n",c1,sqrt(cov11),100*sqrt(cov11)/fabs(c1));
}
fprintf(file," ------------------------------------------------------------\n");
fprintf(file,"\n");
}
int main(int argc,char* argv[]){
double **data=NULL; /* array of values */
size_t size=0; /* length of array vals */
char *splitstring = strdup(" \t");
int o_verbose=0;
int o_model=0;
int o_output=0;
int o_weight=0;
/* COMMAND LINE PROCESSING */
/* variables for reading command line options */
/* ------------------------------------------ */
int opt;
/* ------------------------------------------ */
/* read the command line */
while((opt=getopt_long(argc,argv,"v:hF:M:O:w",gb_long_options, &gb_option_index))!=EOF){
if(opt==0){
gbutils_header(argv[0],stdout);
exit(0);
}
else if(opt=='?'){
fprintf(stderr,"option %c not recognized\n",optopt);
return(-1);
}
else if(opt=='F'){
/* set the fields separator string*/
free(splitstring);
splitstring = strdup(optarg);
}
else if(opt=='M'){
/* set the method to use*/
o_model = atoi(optarg);
}
else if(opt=='O'){
/* set the type of output */
o_output = atoi(optarg);
}
else if(opt=='w'){
/* set third columns as weight */
o_weight = 1;
}
else if(opt=='v'){
/* set verbosity level*/
o_verbose = atoi(optarg);
}
else if(opt=='h'){
/* print short help */
fprintf(stdout,"Linear regression. Data provided should have independent variable on the\n");
fprintf(stdout,"1st column and dependent variable on the 2nd. Standard error on dependent\n");
fprintf(stdout,"variable can be provided on the 3rd column.\n");
fprintf(stdout,"\nUsage: %s [options]\n\n",argv[0]);
fprintf(stdout,"Options:\n");
fprintf(stdout," -M the linear regression model (default 0)\n");
fprintf(stdout," 0 with estimated intercept \n");
fprintf(stdout," 1 with zero intercept \n");
fprintf(stdout," -O the type of output (default 0)\n");
fprintf(stdout," 0 regression coefficients \n");
fprintf(stdout," 1 regression coefficients and errors \n");
fprintf(stdout," 2 x, fitted y, error on y, residual \n");
fprintf(stdout," -w consider provided standard errors on y \n");
fprintf(stdout," -v verbosity level (default 0)\n");
fprintf(stdout," 0 just output \n");
fprintf(stdout," 1 commented headings \n");
fprintf(stdout," 2 model details \n");
fprintf(stdout," -F specify the input fields separators (default \" \\t\")\n");
fprintf(stdout," -h print this help\n");
exit(0);
}
}
/* END OF COMMAND LINE PROCESSING */
/* initialize global variables */
initialize_program(argv[0]);
switch(o_model){
case 0 :
{
double c0,c1,cov00,cov01,cov11,sumsq;
if(!o_weight){
load2(&data,&size,0,splitstring);
gsl_fit_linear (data[0],1, data[1],1, size,
&c0,&c1,&cov00,&cov01,&cov11,&sumsq);
}
else {
size_t i;
load3(&data,&size,0,splitstring);
for(i=0;i<size;i++){ /* build weights from std. dev. */
const double dtmp1=data[2][i];
data[2][i]=1./(dtmp1*dtmp1);
}
gsl_fit_wlinear (data[0],1,data[2],1,data[1],1, size,
&c0,&c1,&cov00,&cov01,&cov11,&sumsq);
}
/* in case, print verbose output */
if(o_verbose>1)
print_linear_verbose(stdout,o_weight,o_model,size,c0,c1,cov00,cov01,cov11,sumsq);
switch(o_output){
case 0:
/* +++++++++++++++++++++++++++++++ */
if(o_verbose>0){
fprintf(stdout,EMPTY_SEP,"#c0");
fprintf(stdout,EMPTY_NL,"c1");
}
/* +++++++++++++++++++++++++++++++ */
fprintf(stdout,FLOAT_SEP,c0);
fprintf(stdout,FLOAT_NL,c1);
break;
case 1:
/* +++++++++++++++++++++++++++++++ */
if(o_verbose>0){
fprintf(stdout,EMPTY_SEP,"#c0");
fprintf(stdout,EMPTY_SEP,"+/-");
fprintf(stdout,EMPTY_SEP,"c1");
fprintf(stdout,EMPTY_NL,"+/-");
}
/* +++++++++++++++++++++++++++++++ */
fprintf(stdout,FLOAT_SEP,c0);
fprintf(stdout,FLOAT_SEP,sqrt(cov00));
fprintf(stdout,FLOAT_SEP,c1);
fprintf(stdout,FLOAT_NL,sqrt(cov11));
break;
case 2:
{
size_t i;
double y,y_err;
/* +++++++++++++++++++++++++++++++ */
if(o_verbose>0){
fprintf(stdout,EMPTY_SEP,"#c0");
fprintf(stdout,EMPTY_SEP,"+/-");
fprintf(stdout,EMPTY_SEP,"c1");
fprintf(stdout,EMPTY_NL,"+/-");
fprintf(stdout,"#");
fprintf(stdout,FLOAT_SEP,c0);
fprintf(stdout,FLOAT_SEP,sqrt(cov00));
fprintf(stdout,FLOAT_SEP,c1);
fprintf(stdout,FLOAT_NL,sqrt(cov11));
}
/* +++++++++++++++++++++++++++++++ */
for(i=0;i<size;i++){
gsl_fit_linear_est (data[0][i], c0,c1,cov00,cov01,cov11,&y,&y_err);
fprintf(stdout,FLOAT_SEP,data[0][i]);
fprintf(stdout,FLOAT_SEP,y);
fprintf(stdout,FLOAT_SEP,y_err);
fprintf(stdout,FLOAT_NL,data[1][i]-y);
}
}
}
}
break;
case 1 :
{
double c1,cov11,sumsq;
if(!o_weight){
load2(&data,&size,0,splitstring);
gsl_fit_mul (data[0],1,data[1],1, size,&c1,&cov11,&sumsq);
}
else {
size_t i;
load3(&data,&size,0,splitstring);
for(i=0;i<size;i++){ /* build weights from std. dev. */
const double dtmp1=data[2][i];
data[2][i]=1./(dtmp1*dtmp1);
}
gsl_fit_wmul (data[0],1,data[2],1,data[1],1, size,&c1,&cov11,&sumsq);
}
/* in case, print verbose output */
if(o_verbose>1)
print_linear_verbose(stdout,o_weight,o_model,size,0.0,c1,0.0,0.0,cov11,sumsq);
switch(o_output){
case 0:
/* +++++++++++++++++++++++++++++++ */
if(o_verbose>0)
fprintf(stdout,EMPTY_NL,"#c1");
/* +++++++++++++++++++++++++++++++ */
fprintf(stdout,FLOAT_NL,c1);
break;
case 1:
/* +++++++++++++++++++++++++++++++ */
if(o_verbose>0){
fprintf(stdout,EMPTY_SEP,"#c1");
fprintf(stdout,EMPTY_NL,"+/-");
}
/* +++++++++++++++++++++++++++++++ */
fprintf(stdout,FLOAT_SEP,c1);
fprintf(stdout,FLOAT_NL,sqrt(cov11));
break;
case 2:
{
size_t i;
double y,y_err;
/* +++++++++++++++++++++++++++++++ */
if(o_verbose>0){
fprintf(stdout,EMPTY_SEP,"#c1");
fprintf(stdout,EMPTY_NL,"+/-");
}
/* +++++++++++++++++++++++++++++++ */
fprintf(stdout,FLOAT_SEP,c1);
fprintf(stdout,FLOAT_NL,sqrt(cov11));
for(i=0;i<size;i++){
gsl_fit_mul_est (data[0][i],c1,cov11,&y,&y_err);
fprintf(stdout,FLOAT_SEP,data[0][i]);
fprintf(stdout,FLOAT_SEP,y);
fprintf(stdout,FLOAT_SEP,y_err);
fprintf(stdout,FLOAT_NL,data[1][i]-y);
}
}
}
}
break;
}
return 0;
}
|
