int
Clusters::
reclusterNeyEssen()
{
cerr << "Calculating MLE for prior probabilities" << endl;
vector<double> prior(numberClasses,0.0l);
double xxx = 1.0l/(double)numberTypes;
for (int i = 0; i < numberTypes; i++){
int c = classVector[i];
prior[c] += xxx;
}
for (int i = 0; i < numberClasses;i++)
cerr << i << " " << prior[i] << endl;
if (numberStates > 0){
cerr << "Training all the HMMs" << endl;
for (int c = 0; c < numberClasses; c++){
// cerr << "Training HMM " << c << endl;
HMM* hmmPtr = hmms[c];
HMM* newHmmPtr = new HMM(numberStates, alphabetSize);
for (int i = 0; i < numberTypes; i++){
if (classVector[i] == c){
// then this word is in the right class
// so train it on word i
const string & word = *(corpus.wordArray[i]);
vector<int> v;
hmmPtr->convertString(word,v);
// FIXME
double weight = 1.0l;
if (USE_TRUE_WEIGHT){
weight = corpus.countArray[i];
}
hmmPtr->emSingle(*newHmmPtr, weight, v);
}
}
newHmmPtr->normalise();
hmms[c] = newHmmPtr;
delete hmmPtr;
}
}
int something = 0;
for (int i = 0; i < numberTypes; i++){
// cerr << "Word " << i;
int w = sortedWords[i];
//cerr << *(corpus.wordArray[w]) << endl;
if (counts[w] > FREQ_CUTOFF){
//cerr << "Doing " << w << endl;
if (bestCluster(w, prior)){
something++;
}
}
}
return something;
}
/**
* Ejecuta la metaheurística con los datos dados.
*
* @param type Si se utilizará una función objetivo de
* maximización o de minimización.
*/
void Kmeans::run(int type){
int i;
////////////////////////////
// Inicializaciones.
//Inicializaciones del mejor según el tipo de problema.
float best;
if(initialized){
switch(type){
case T_MAX:
//Maximización.
best = 0.0;
break;
default:
//Minimización.
best = numeric_limits<float>::infinity();
}
}else{
best = bestFO;
}
int count = 0;
////////////////////////////
// Algoritmo K-means.
while(count < REPS){
//////////////////////////////
// Reasignación y evaluación.
//Asignar elementos a cada cluster.
for(i = 0; i < N; ++i)
solution[0][i] = bestCluster(0, i);
//Renombra los clusters y recalcula los centroides.
renamer(0, &Ks[0], size);
//Evalua la función objetivo.
switch(type){
case T_MAX:
of[0] = foMax(0, Ks[0], metric);
break;
default:
of[0] = foMin(0, Ks[0], metric);
}
////////////////////////////
// Actualización del mejor.
updateBetter(0, &best, &count, type);
}
}
