void LDA::gibbs(int K,double alpha,double beta) {
this->K=K;
this->alpha=alpha;
this->beta=beta;
if(SAMPLE_LAG >0) {
thetasum = MatrixXd(documentsSize(),K);
phisum= MatrixXd(K,V);
numstats=0;
}
initialState(K);
for(int i=0; i<ITERATIONS; i++) {
for(int m=0; m<z.rows(); m++) {
for(int n=0; n<z.cols(); n++) {
z(m,n)=sampleFullConditional(m,n);
}
}
if(i > BURN_IN && (i%THIN_INTERVAL==0)) {
dispcol++;
}
if ((i > BURN_IN) && (SAMPLE_LAG > 0) && (i % SAMPLE_LAG == 0)) {
updateParams();
if (i % THIN_INTERVAL != 0)
dispcol++;
}
if (dispcol >= 100) {
dispcol = 0;
}
}
}
MatrixXd LDA::getTheta(void) {
if(SAMPLE_LAG>0) {
return thetasum*(1.0/numstats);
}
else {
MatrixXd theta(documentsSize(),K);
for(int m=0; m<documentsSize(); m++) {
// theta.row(m)=((nd.row(m).cast<double>().array()+alpha)/(ndsum.cast<double>().array()+K*alpha)).matrix();
for(int k=0; k<K; k++) {
theta(m,k)=(nd(m,k)+alpha)/(ndsum(m)+K*alpha);
}
}
return theta;
}
}
void LDA::updateParams(void) {
for(int m=0; m<documentsSize(); m++) {
for(int k=0; k<K; k++) {
thetasum(m,k)+=(nd(m,k)+alpha)/(ndsum(m)+K*alpha);
}
}
for(int k=0; k<K; k++) {
for(int w=0; w<V; w++) {
phisum(k,w)+= (nw(w,k)+beta)/(nwsum(k)+V*beta);
}
}
numstats++;
}
void LDA::initialState(int K) {
int M=documentsSize(),N=(*documents).cols();
nw = MatrixXi::Zero(V,K);
nd = MatrixXi::Zero(M,K);
nwsum = VectorXi::Zero(K);
ndsum = VectorXi(M);
z=(MatrixXd::Random(M,N).cwiseAbs()*K).cast<int>();
ndsum.fill(N);
for(int i=0; i<M; i++) {
for(int j=0; j<N; j++) {
nw((*documents)(i,j),z(i,j))++;
nd(i,z(i,j))++;
nwsum(z(i,j))++;
}
}
}
int FmDriveDetailsContent::getDataSizeByTraversePath( const QString &driveName,
QList<FmDriveDetailsSize*> &detailsSizeList, volatile bool *isStopped )
{
qint64 imageSize( 0 );
qint64 soundSize( 0 );
qint64 midpJavaSize( 0 );
qint64 nativeAppsSize( 0 );
qint64 videoSize( 0 );
qint64 documentsSize( 0 );
FmFileTypeRecognizer fileTypeRecognizer;
QList<QDir> dirs;
dirs.append( QDir( driveName ) );
// traverse the whole drive
while (!dirs.isEmpty()) {
QDir::Filters filter = QDir::NoDotAndDotDot | QDir::AllEntries;
// do not summarize system and hidden files, these size will go into others category
// if( isSysHiddenIncluded ) {
// filter = filter | QDir::Hidden | QDir::System;
// }
QFileInfoList infoList = dirs.first().entryInfoList( filter );
for ( QFileInfoList::const_iterator it = infoList.begin(); it != infoList.end(); ++it ) {
if ( *isStopped ){
return FmErrCancel;
}
if ( it->isFile() ) {
FmFileTypeRecognizer::FileType fileType =
fileTypeRecognizer.getType( it->absoluteFilePath() );
switch ( fileType )
{
case FmFileTypeRecognizer::FileTypeImage:
imageSize += it->size();
break;
case FmFileTypeRecognizer::FileTypeTone:
soundSize += it->size();
break;
case FmFileTypeRecognizer::FileTypeJava:
midpJavaSize += it->size();
break;
case FmFileTypeRecognizer::FileTypeSisx:
nativeAppsSize += it->size();
break;
case FmFileTypeRecognizer::FileTypeVideo:
videoSize += it->size();
break;
case FmFileTypeRecognizer::FileTypeText:
documentsSize += it->size();
break;
default:
// do not need handle other type
break;
}
}
else if ( it->isDir() ) {
dirs.append( QDir( it->absoluteFilePath() ) );
}
}
dirs.removeFirst();
}
// store result to detailsSizeList.
detailsSizeList.append( new FmDriveDetailsSize( FmDriveDetailsSize::ETypeImages, imageSize ) ) ;
detailsSizeList.append( new FmDriveDetailsSize( FmDriveDetailsSize::ETypeSoundFiles, soundSize ) );
detailsSizeList.append( new FmDriveDetailsSize( FmDriveDetailsSize::ETypeMidpJava, midpJavaSize ) );
detailsSizeList.append( new FmDriveDetailsSize( FmDriveDetailsSize::ETypeNativeApps, nativeAppsSize ) );
detailsSizeList.append( new FmDriveDetailsSize( FmDriveDetailsSize::ETypeVideos, videoSize ) );
detailsSizeList.append( new FmDriveDetailsSize( FmDriveDetailsSize::ETypeDocuments, documentsSize ) );
return FmErrNone;
}
