int main(int argc, char* argv[]) {
// open file
FILE *f = fopen(argv[1],"r");
// Throw an error if file not found
if (f == NULL) {
printf("Sorry, file not found.\n");
}
// Take the first line which is the count of number of lines
int n;
fscanf(f, "%d", &n);
if (n > 4097) {
printf("Sorry, file too large, cannot be more than 4097 lines");
}
if (n < 2) {
printf("Sorry, file too short, cannot be less than 2 lines");
}
// Store the values from file in array
int values[n];
int i;
for (i = 0; i < n; i++) {
fscanf(f, "%d", &values[i]);
}
// Sort the values in ascending order
qsort(values, n, sizeof(int), comparator);
// Calculate the mean
printf ("Mean = %0.2f\n", meanFunc(values, n));
// Calculate the median
printf ("Median = %0.2f\n", medianFunc(values, n));
// Calculate the mode
printf("Mode = %0.2f\n", modeFunc(values, n));
// Find & print min value
printf ("Min = %0.2f\n", minFunc(values));
// Find & print max value
printf ("Max = %0.2f\n", maxFunc(values, n));
return 0;
}
/* @returns: 0 -> success
* -1 -> inconsistent lower/upper specification
* -2 -> can't open pSampleInputFile
* -3 -> can't allocate enough memory
* -4 -> can't open pOutputParameters
* -5 -> can't open pInputParameters
* -6 -> can't open pScaledSampleOutputFile
* -7 -> cannot use read and store of parameters simultaneously
*
*/
int SVM::scaleSamples(char *pSampleInputFile, char* pScaledSampleOutputFile, double dLowerBound, double dUpperBound, char* pInputParameters, char* pOutputParameters)
{
char *line = NULL;
int max_line_len = 1024;
double lower=-1.0,upper=1.0,y_lower,y_upper;
int y_scaling = 0;
double *feature_max;
double *feature_min;
double y_max = -DBL_MAX;
double y_min = DBL_MAX;
int max_index;
long int num_nonzeros = 0;
int i,index;
FILE *fp, *fp_restore = NULL;
if(!(upper > lower) )
{
fprintf(stderr,"inconsistent lower/upper specification\n");
return -1;
}
if(pInputParameters != NULL && pOutputParameters != NULL)
{
fprintf(stderr,"cannot use read and store of parameters simultaneously\n");
return -7;
}
fp=fopen(pSampleInputFile,"r");
if(fp==NULL)
{
fprintf(stderr,"can't open file %s\n", pSampleInputFile);
return -2;
}
line = (char *) malloc(max_line_len*sizeof(char));
/* assumption: min index of attributes is 1 */
/* pass 1: find out max index of attributes */
max_index = 0;
if(pInputParameters != NULL)
{
int idx, c;
fp_restore = fopen(pInputParameters,"r");
if(fp_restore==NULL)
{
fprintf(stderr,"can't open file %s\n", pInputParameters);
return -5;
}
c = fgetc(fp_restore);
if(c == 'y')
{
line = this->read_line(fp_restore, 1024);
line = this->read_line(fp_restore, 1024);
line = this->read_line(fp_restore, 1024);
}
line = this->read_line(fp_restore, 1024);
line = this->read_line(fp_restore, 1024);
while(fscanf(fp_restore,"%d %*f %*f\n",&idx) == 1)
max_index = max(idx,max_index);
rewind(fp_restore);
}
while((line = this->read_line(fp, 1024))!=NULL)
{
char *p=line;
SKIP_TARGET
while(sscanf(p,"%d:%*f",&index)==1)
{
max_index = max(max_index, index);
SKIP_ELEMENT
num_nonzeros++;
}
}
rewind(fp);
feature_max = (double *)malloc((max_index+1)* sizeof(double));
feature_min = (double *)malloc((max_index+1)* sizeof(double));
if(feature_max == NULL || feature_min == NULL)
{
fprintf(stderr,"can't allocate enough memory\n");
return -3;
}
for(i=0;i<=max_index;i++)
{
feature_max[i]=-DBL_MAX;
feature_min[i]=DBL_MAX;
}
/* pass 2: find out min/max value */
while((line = this->read_line(fp, 1024))!=NULL)
{
char *p=line;
int next_index=1;
double target;
double value;
sscanf(p,"%lf",&target);
y_max = max(y_max,target);
y_min = min(y_min,target);
SKIP_TARGET
while(sscanf(p,"%d:%lf",&index,&value)==2)
{
for(i=next_index;i<index;i++)
{
feature_max[i]=maxFunc(feature_max[i],0);
feature_min[i]=minFunc(feature_min[i],0);
}
feature_max[index]=maxFunc(feature_max[index],value);
feature_min[index]=minFunc(feature_min[index],value);
SKIP_ELEMENT
next_index=index+1;
}
for(i=next_index;i<=max_index;i++)
{
feature_max[i]=maxFunc(feature_max[i],0);
feature_min[i]=minFunc(feature_min[i],0);
}
}
rewind(fp);
/* pass 2.5: save/restore feature_min/feature_max */
if(pInputParameters != NULL)
{
/* fp_restore rewinded in finding max_index */
int idx, c;
double fmin, fmax;
if((c = fgetc(fp_restore)) == 'y')
{
fscanf(fp_restore, "%lf %lf\n", &y_lower, &y_upper);
fscanf(fp_restore, "%lf %lf\n", &y_min, &y_max);
y_scaling = 1;
}
else
ungetc(c, fp_restore);
if (fgetc(fp_restore) == 'x') {
fscanf(fp_restore, "%lf %lf\n", &lower, &upper);
while(fscanf(fp_restore,"%d %lf %lf\n",&idx,&fmin,&fmax)==3)
{
if(idx<=max_index)
{
feature_min[idx] = fmin;
feature_max[idx] = fmax;
}
}
}
fclose(fp_restore);
}
if(pOutputParameters != NULL)
{
FILE *fp_save = fopen(pOutputParameters,"w");
if(fp_save==NULL)
{
fprintf(stderr,"can't open file %s\n", pOutputParameters);
exit(1);
}
if(y_scaling)
{
fprintf(fp_save, "y\n");
fprintf(fp_save, "%.16g %.16g\n", y_lower, y_upper);
fprintf(fp_save, "%.16g %.16g\n", y_min, y_max);
}
fprintf(fp_save, "x\n");
fprintf(fp_save, "%.16g %.16g\n", lower, upper);
for(i=1;i<=max_index;i++)
{
if(feature_min[i]!=feature_max[i])
fprintf(fp_save,"%d %.16g %.16g\n",i,feature_min[i],feature_max[i]);
}
fclose(fp_save);
}
/* pass 3: scale */
FILE *fp_scaledOutput = fopen(pScaledSampleOutputFile,"w");
if(fp_scaledOutput==NULL)
{
fprintf(stderr,"can't open file %s\n", pScaledSampleOutputFile);
return -6;
}
while((line = this->read_line(fp,1024))!=NULL)
{
char *p=line;
int next_index=1;
double target;
double value;
sscanf(p,"%lf",&target);
fprintf(fp_scaledOutput, "%s", this->output_target(target, y_scaling, y_lower, y_upper, y_min, y_max));
SKIP_TARGET
while(sscanf(p,"%d:%lf",&index,&value)==2)
{
for(i=next_index;i<index;i++)
fprintf(fp_scaledOutput, "%s", this->output(i,0, feature_min, feature_max, lower, upper));
fprintf(fp_scaledOutput, "%s", this->output(index, value, feature_min, feature_max, lower, upper));
SKIP_ELEMENT
next_index=index+1;
}
for(i=next_index;i<=max_index;i++)
fprintf(fp_scaledOutput, "%s", this->output(i, 0, feature_min, feature_max, lower, upper));
fprintf(fp_scaledOutput, "\n");
}
/*
if (new_num_nonzeros > num_nonzeros)
fprintf(stderr,
"Warning: original #nonzeros %ld\n"
" new #nonzeros %ld\n"
"Use -l 0 if many original feature values are zeros\n",
num_nonzeros, new_num_nonzeros);
*/
free(line);
free(feature_max);
free(feature_min);
fclose(fp);
fclose(fp_scaledOutput);
return 0;
}