int DepPipe::writeInstance(FILE *featFile, DepInstance *pInstance)
{
// cerr << endl;
int instanceLength = pInstance->size();
for(int w1 = 0; w1 < instanceLength; w1++) {
for(int w2 = w1+1; w2 < instanceLength; w2++) {
for(int ph = 0; ph < 2; ph++) {
bool attR = ph == 0 ? true : false;
FeatureVec prodFV;
addArcFeature(pInstance,w1,w2,attR,prodFV);
vector<int> vecKeys;
prodFV.getKeys(vecKeys);
// cerr << vecKeys.size() << " ";
::writeObject(featFile, vecKeys);
}
}
// cerr << endl;
}
::writeObject(featFile, (int)-1);
if(options.m_isLabeled) {
for(int w1 = 0; w1 < instanceLength; w1++) {
for(int t = 0; t < m_vecTypes.size(); t++) {
const string &type = m_vecTypes[t];
for(int ph = 0; ph < 2; ph++) {
bool attR = ph == 0 ? true : false;
for(int ch = 0; ch < 2; ch++) {
bool child = ch == 0 ? true : false;
FeatureVec prodFV;
addLabelFeature(pInstance, w1, type, child, attR, prodFV);
vector<int> vecKeys;
prodFV.getKeys(vecKeys);
// cerr << vecKeys.size() << " ";
// copy(vecKeys.begin(), vecKeys.end(), ostream_iterator<int>(cerr, " "));
// cerr << endl;
::writeObject(featFile, vecKeys);
}
}
}
// cerr << endl;
}
::writeObject(featFile, int(-2));
}
// exit(0);
writeExtendedFeatures(pInstance,featFile);
vector<int> vecKeys;
pInstance->fv.getKeys(vecKeys);
::writeObject(featFile, vecKeys);
::writeObject(featFile, int(-3));
// cerr << pInstance->actParseTree.size() << endl;
writeObject(featFile, pInstance->actParseTree);
writeObject(featFile, int(-4));
return 0;
}
double MultiClassExp::getConfidence(const FeatureVec& feature_vec, const std::vector<double>& weights_plus, const std::vector<double>& weights_minus)
{
double confidence_plus = std::inner_product(feature_vec.begin(), feature_vec.end(),
weights_plus.begin(), 0.0);
double confidence_minus = std::inner_product(feature_vec.begin(), feature_vec.end(),
weights_minus.begin(), 0.0);
double confidence = confidence_plus-confidence_minus;
return confidence;
}
/*the document is index:value format because there is less features,centorid is array format for quick access
*now it's changed into cosine!2011-07-22
* */
float KNN::InnerProduct(FeatureVec & document,vector<float> & centorid, int catIndex)
{
/*part of cosine*/
float similarity=0.0;
float finalScore=0.0;
float totalDoc=0.0;
FeatVecIte it;
for(it=document.begin();it!=document.end();it++)
{
similarity+=((*it).weight*centorid[(*it).index]);//direct access the vector
}
//float docLen=inner_product(centorid.begin(),centorid.end(),centorid.begin(),0.0);//tip:可以提前计算好。@zhqm
return similarity;
}
int KNN::MaxProbCat(FeatureVec & document,vector<vector<float> > & Centoroid,float & score)
{
float MaxScoreNow=0.0;
int MaxCatIndex=-1;
int currentCatIndex=0;
float docDotProduct=0.0;
FeatVecIte dit;
for(dit=document.begin();dit!=document.end();dit++)
{
docDotProduct+=(*dit).weight*(*dit).weight; //tip:@zhqm totalDoc +=
}
docDotProduct = sqrt(docDotProduct);
if (docDotProduct == 0){
score = 0.0;
return MaxCatIndex;
}
vector<vector<float> >::iterator it;
for(it=Centoroid.begin();it!=Centoroid.end();it++,currentCatIndex++)
{
float score=InnerProduct(document,(*it), currentCatIndex);
score=(float)score/(docDotProduct * m_centoroidFeatureDotProduct.at(currentCatIndex));
//cout<<"cat="<<currentCatIndex<<"\tscore="<<score<<endl;
if(score>MaxScoreNow)//a better class
{
MaxScoreNow=score;
MaxCatIndex=currentCatIndex;
}
}
score=MaxScoreNow;
return MaxCatIndex;
}
int DepPipe::readInstance(FILE *featFile, int length,
MultiArray<FeatureVec> &fvs,
MultiArray<double> &probs,
MultiArray<FeatureVec> &fvs_trips,
MultiArray<double> &probs_trips,
MultiArray<FeatureVec> &fvs_sibs,
MultiArray<double> &probs_sibs,
MultiArray<FeatureVec> &nt_fvs,
MultiArray<double> &nt_probs,
FeatureVec &fv,
string &actParseTree,
const Parameter ¶ms)
{
// cerr << "read instance" << endl;
// cerr << endl;
vector<unsigned int> fvs_dim;
unsigned int fvs_pos;
MyVector<int> vecKeys;
// Get production crap.
for(int w1 = 0; w1 < length; w1++) {
int w2 = w1 + 1;
if (w2 >= length) continue;
fvs.setDemisionVal(fvs_dim, w1, w2, 0);
fvs.getElement(fvs_dim, fvs_pos);
for(; w2 < length; w2++) {
for(int ph = 0; ph < 2; ph++) {
FeatureVec &prodFV = fvs.getElement(fvs_pos);
::readObject(featFile, vecKeys);
// cerr << vecKeys.size() << " ";
prodFV.setKeys(vecKeys);
probs.getElement(fvs_pos) = params.getScore(prodFV);
++fvs_pos;
}
}
// cerr << endl;
}
int last;
::readObject(featFile, last);
if(last != -1) { cerr << "DepPipe::readInstance() Error reading file. -1" << endl; return -1; }
// cerr << endl;
if(options.m_isLabeled) {
vector<unsigned int> nt_dim(4);
unsigned int nt_pos;
nt_fvs.setDemisionVal(nt_dim, 0, 0, 0, 0);
nt_fvs.getElement(nt_dim, nt_pos);
for(int w1 = 0; w1 < length; w1++) {
for(int t = 0; t < m_vecTypes.size(); t++) {
const string &type = m_vecTypes[t];
for(int ph = 0; ph < 2; ph++) {
for(int ch = 0; ch < 2; ch++) {
FeatureVec &prodFV = nt_fvs.getElement(nt_pos);
::readObject(featFile, vecKeys);
prodFV.setKeys(vecKeys);
// cerr << vecKeys.size() << " ";
nt_probs.getElement(nt_pos) = params.getScore(prodFV);
++nt_pos;
}
}
}
// cerr << endl;
}
::readObject(featFile, last);
if(last != -2) { cerr << "DepPipe::readInstance() Error reading file.-2 vs. " << last << endl; return -1; }
}
// exit(0);
::readObject(featFile, vecKeys);
fv.setKeys(vecKeys);
::readObject(featFile, last);
if(last != -3) { cerr << "DepPipe::readInstance() Error reading file.-3 vs. " << last << endl; return -1; }
MyVector<char> my_str;
::readObject(featFile, my_str);
actParseTree = my_str.begin();
// cerr << actParseTree << endl;
// cerr << actParseTree.size() << endl;
::readObject(featFile, last);
if(last != -4) { cerr << "DepPipe::readInstance() Error reading file.-4 vs. " << last << endl; return -1; }
return 0;
}
void DepPipe::add(const string &feat, double val, FeatureVec &fv) {
int num = m_featAlphabet.lookupIndex(feat);
if (num >= 0) {
fv.add(num, val);
}
}