SAMPLE read_struct_examples(char *file, STRUCT_LEARN_PARM *sparm)
{
/* Reads training examples and returns them in sample. The number of
examples must be written into sample.n */
SAMPLE sample; /* sample */
EXAMPLE *examples;
long n; /* number of examples */
DOC **docs; /* examples in original SVM-light format */
double *target;
long totwords, i, num_classes = 0;
/* Using the read_documents function from SVM-light */
read_documents(file, &docs, &target, &totwords, &n);
examples = (EXAMPLE *)my_malloc(sizeof(EXAMPLE)*n);
for( i=0; i<n; i++ ) /* find highest class label */
if(num_classes < (target[i]+0.1))
num_classes = target[i]+0.1;
for( i=0; i<n; i++ ) /* make sure all class labels are positive */
if(target[i]<1) {
printf("\nERROR: The class label '%lf' of example number %ld is not greater than '1'!\n",target[i],i+1);
exit(1);
}
for( i=0; i<n; i++ ) { /* copy docs over into new datastructure */
examples[i].x.doc = docs[i];
examples[i].y.classlabel = target[i]+0.1;
examples[i].y.scores = NULL;
examples[i].y.num_classes = num_classes;
}
free(target);
free(docs);
sample.n = n;
sample.examples = examples;
if(struct_verbosity>=0)
printf(" (%d examples) ",sample.n);
return(sample);
}
int _svm_learn (int argc, char* argv[])
{
char docfile[200]; /* file with training examples */
char modelfile[200]; /* file for resulting classifier */
char restartfile[200]; /* file with initial alphas */
DOC **docs; /* training examples */
long totwords,totdoc,i;
double *target;
double *alpha_in=NULL;
KERNEL_CACHE *kernel_cache;
LEARN_PARM learn_parm;
KERNEL_PARM kernel_parm;
MODEL *model=(MODEL *)my_malloc(sizeof(MODEL));
HIDEO_ENV *hideo_env=create_env();
model->td_pred=NULL;
model->n_td_pred=0;
_read_input_parameters(argc,argv,docfile,modelfile,restartfile,&verbosity,
&learn_parm,&kernel_parm);
read_documents(docfile,&docs,&target,&totwords,&totdoc);
if(restartfile[0]) alpha_in=read_alphas(restartfile,totdoc);
if(kernel_parm.kernel_type == LINEAR) { /* don't need the cache */
kernel_cache=NULL;
}
else {
/* Always get a new kernel cache. It is not possible to use the
same cache for two different training runs */
kernel_cache=kernel_cache_init(totdoc,learn_parm.kernel_cache_size);
}
if(learn_parm.type == CLASSIFICATION) {
svm_learn_classification(docs,target,totdoc,totwords,&learn_parm,
&kernel_parm,kernel_cache,model,alpha_in,hideo_env);
}
else if(learn_parm.type == REGRESSION) {
svm_learn_regression(docs,target,totdoc,totwords,&learn_parm,
&kernel_parm,&kernel_cache,model,hideo_env);
}
else if(learn_parm.type == RANKING) {
svm_learn_ranking(docs,target,totdoc,totwords,&learn_parm,
&kernel_parm,&kernel_cache,model,hideo_env);
}
else if(learn_parm.type == OPTIMIZATION) {
svm_learn_optimization(docs,target,totdoc,totwords,&learn_parm,
&kernel_parm,kernel_cache,model,alpha_in,hideo_env);
}
if(kernel_cache) {
/* Free the memory used for the cache. */
kernel_cache_cleanup(kernel_cache);
}
/* Warning: The model contains references to the original data 'docs'.
If you want to free the original data, and only keep the model, you
have to make a deep copy of 'model'. */
/* deep_copy_of_model=copy_model(model); */
write_model(modelfile,model);
free(alpha_in);
free_model(model,0);
for(i=0;i<totdoc;i++)
free_example(docs[i],1);
free(docs);
free(target);
free_env(hideo_env);
return(0);
}
int main (int argc, char* argv[])
{
DOC *docs; /* training examples */
long max_docs,max_words_doc;
long totwords,totdoc,ll,i;
long kernel_cache_size;
double *target;
KERNEL_CACHE kernel_cache;
LEARN_PARM learn_parm;
KERNEL_PARM kernel_parm;
MODEL model;
read_input_parameters(argc,argv,docfile,modelfile,&verbosity,
&kernel_cache_size,&learn_parm,&kernel_parm);
if(verbosity>=1) {
printf("Scanning examples..."); fflush(stdout);
}
nol_ll(docfile,&max_docs,&max_words_doc,&ll); /* scan size of input file */
max_words_doc+=10;
ll+=10;
max_docs+=2;
if(verbosity>=1) {
printf("done\n"); fflush(stdout);
}
docs = (DOC *)my_malloc(sizeof(DOC)*max_docs); /* feature vectors */
target = (double *)my_malloc(sizeof(double)*max_docs); /* target values */
//printf("\nMax docs: %ld, approximated number of feature occurences %ld, maximal length of a line %ld\n\n",max_docs,max_words_doc,ll);
read_documents(docfile,docs,target,max_words_doc,ll,&totwords,&totdoc,&kernel_parm);
printf("\nNumber of examples: %ld, linear space size: %ld\n\n",totdoc,totwords);
//if(kernel_parm.kernel_type==5) totwords=totdoc; // The number of features is proportional to the number of parse-trees, i.e. totdoc
// or should we still use totwords to approximate svm_maxqpsize for the Tree Kernel (see hideo.c) ???????
if(kernel_parm.kernel_type == LINEAR) { /* don't need the cache */
if(learn_parm.type == CLASSIFICATION) {
svm_learn_classification(docs,target,totdoc,totwords,&learn_parm,
&kernel_parm,NULL,&model);
}
else if(learn_parm.type == REGRESSION) {
svm_learn_regression(docs,target,totdoc,totwords,&learn_parm,
&kernel_parm,NULL,&model);
}
else if(learn_parm.type == RANKING) {
svm_learn_ranking(docs,target,totdoc,totwords,&learn_parm,
&kernel_parm,NULL,&model);
}
}
else {
if(learn_parm.type == CLASSIFICATION) {
/* Always get a new kernel cache. It is not possible to use the
same cache for two different training runs */
kernel_cache_init(&kernel_cache,totdoc,kernel_cache_size);
svm_learn_classification(docs,target,totdoc,totwords,&learn_parm,
&kernel_parm,&kernel_cache,&model);
/* Free the memory used for the cache. */
kernel_cache_cleanup(&kernel_cache);
}
else if(learn_parm.type == REGRESSION) {
/* Always get a new kernel cache. It is not possible to use the
same cache for two different training runs */
kernel_cache_init(&kernel_cache,2*totdoc,kernel_cache_size);
svm_learn_regression(docs,target,totdoc,totwords,&learn_parm,
&kernel_parm,&kernel_cache,&model);
/* Free the memory used for the cache. */
kernel_cache_cleanup(&kernel_cache);
}
else if(learn_parm.type == RANKING) {
printf("Learning rankings is not implemented for non-linear kernels in this version!\n");
exit(1);
}
else if(learn_parm.type == PERCEPTRON) {
perceptron_learn_classification(docs,target,totdoc,totwords,&learn_parm,
&kernel_parm,&kernel_cache,&model,modelfile);
}
else if(learn_parm.type == PERCEPTRON_BATCH) {
batch_perceptron_learn_classification(docs,target,totdoc,totwords,&learn_parm,
&kernel_parm,kernel_cache_size,&model);
}
}
/* Warning: The model contains references to the original data 'docs'.
If you want to free the original data, and only keep the model, you
have to make a deep copy of 'model'. */
write_model(modelfile,&model);
free(model.supvec);
free(model.alpha);
free(model.index);
for(i=0;i<totdoc;i++){
freeExample(&docs[i]);
}
free(docs);
free(target);
return(0);
}