public static SparseVector <double> ComputeCentroid(IEnumerable <int> vecIdxList, IUnlabeledExampleCollection <SparseVector <double> > dataset, CentroidType type)
{
Utils.ThrowException(vecIdxList == null ? new ArgumentNullException("vecIdxList") : null);
Utils.ThrowException(dataset == null ? new ArgumentNullException("dataset") : null);
Dictionary <int, double> tmp = new Dictionary <int, double>();
int vecCount = 0;
foreach (int vecIdx in vecIdxList)
{
Utils.ThrowException((vecIdx < 0 || vecIdx >= dataset.Count) ? new ArgumentValueException("vecIdxList") : null);
SparseVector <double> vec = dataset[vecIdx];
foreach (IdxDat <double> item in vec)
{
if (tmp.ContainsKey(item.Idx))
{
tmp[item.Idx] += item.Dat;
}
else
{
tmp.Add(item.Idx, item.Dat);
}
}
vecCount++;
}
//Utils.ThrowException(vecCount == 0 ? new ArgumentValueException("vecIdxList") : null);
if (vecCount == 0)
{
return(new SparseVector <double>());
}
SparseVector <double> centroid = new SparseVector <double>();
switch (type)
{
case CentroidType.Sum:
foreach (KeyValuePair <int, double> item in tmp)
{
centroid.InnerIdx.Add(item.Key);
centroid.InnerDat.Add(item.Value);
}
break;
case CentroidType.Avg:
foreach (KeyValuePair <int, double> item in tmp)
{
centroid.InnerIdx.Add(item.Key);
centroid.InnerDat.Add(item.Value / (double)vecCount);
}
break;
case CentroidType.NrmL2:
double vecLen = 0;
foreach (KeyValuePair <int, double> item in tmp)
{
vecLen += item.Value * item.Value;
}
//Utils.ThrowException(vecLen == 0 ? new InvalidOperationException() : null);
vecLen = Math.Sqrt(vecLen);
if (vecLen > 0)
{
foreach (KeyValuePair <int, double> item in tmp)
{
centroid.InnerIdx.Add(item.Key);
centroid.InnerDat.Add(item.Value / vecLen);
}
}
break;
}
centroid.Sort();
return(centroid);
}
public static SparseVector <double> ComputeCentroidWgt(IEnumerable <Pair <double, SparseVector <double> > > wgtVecList, CentroidType type)
{
Utils.ThrowException(wgtVecList == null ? new ArgumentNullException("wgtVecList") : null);
Dictionary <int, double> tmp = new Dictionary <int, double>();
double wgtSum = 0;
foreach (Pair <double, SparseVector <double> > wgtVec in wgtVecList)
{
Utils.ThrowException(wgtVec.First < 0 || wgtVec.Second == null ? new ArgumentValueException("wgtVecList") : null);
foreach (IdxDat <double> item in wgtVec.Second)
{
if (tmp.ContainsKey(item.Idx))
{
tmp[item.Idx] += wgtVec.First * item.Dat;
}
else
{
tmp.Add(item.Idx, wgtVec.First * item.Dat);
}
}
wgtSum += wgtVec.First;
}
if (wgtSum == 0)
{
return(new SparseVector <double>());
}
SparseVector <double> centroid = new SparseVector <double>();
switch (type)
{
case CentroidType.Sum:
foreach (KeyValuePair <int, double> item in tmp)
{
centroid.InnerIdx.Add(item.Key);
centroid.InnerDat.Add(item.Value);
}
break;
case CentroidType.Avg:
foreach (KeyValuePair <int, double> item in tmp)
{
centroid.InnerIdx.Add(item.Key);
centroid.InnerDat.Add(item.Value / wgtSum);
}
break;
case CentroidType.NrmL2:
double vecLen = 0;
foreach (KeyValuePair <int, double> item in tmp)
{
vecLen += item.Value * item.Value;
}
//Utils.ThrowException(vecLen == 0 ? new InvalidOperationException() : null);
vecLen = Math.Sqrt(vecLen);
if (vecLen > 0)
{
foreach (KeyValuePair <int, double> item in tmp)
{
centroid.InnerIdx.Add(item.Key);
centroid.InnerDat.Add(item.Value / vecLen);
}
}
break;
}
centroid.Sort();
return(centroid);
}
public SparseVector <double> .ReadOnly ComputeCentroid <LblT>(IExampleCollection <LblT, SparseVector <double> .ReadOnly> dataset, CentroidType type)
{
Utils.ThrowException(dataset == null ? new ArgumentNullException("dataset") : null);
Dictionary <int, double> tmp = new Dictionary <int, double>();
double wgt_sum = 0;
foreach (Pair <double, int> wgt_vec in m_items)
{
Utils.ThrowException((wgt_vec.Second < 0 || wgt_vec.Second >= dataset.Count) ? new IndexOutOfRangeException("Items (dataset index)") : null);
foreach (IdxDat <double> item in dataset[wgt_vec.Second].Example)
{
if (tmp.ContainsKey(item.Idx))
{
tmp[item.Idx] += wgt_vec.First * item.Dat;
}
else
{
tmp.Add(item.Idx, wgt_vec.First * item.Dat);
}
}
wgt_sum += wgt_vec.First;
}
Utils.ThrowException(wgt_sum == 0 ? new ArgumentValueException("Items (weights)") : null);
SparseVector <double> centroid = new SparseVector <double>();
switch (type)
{
case CentroidType.Sum:
foreach (KeyValuePair <int, double> item in tmp)
{
centroid.InnerIdx.Add(item.Key);
centroid.InnerDat.Add(item.Value);
}
break;
case CentroidType.Avg:
foreach (KeyValuePair <int, double> item in tmp)
{
centroid.InnerIdx.Add(item.Key);
centroid.InnerDat.Add(item.Value / wgt_sum);
}
break;
case CentroidType.NrmL2:
double vec_len = 0;
foreach (KeyValuePair <int, double> item in tmp)
{
vec_len += item.Value * item.Value;
}
Utils.ThrowException(vec_len == 0 ? new InvalidOperationException() : null);
vec_len = Math.Sqrt(vec_len);
foreach (KeyValuePair <int, double> item in tmp)
{
centroid.InnerIdx.Add(item.Key);
centroid.InnerDat.Add(item.Value / vec_len);
}
break;
}
centroid.Sort();
return(centroid);
}
public SparseVector <double> ComputeCentroid(IUnlabeledExampleCollection <SparseVector <double> > dataset, CentroidType type)
{
return(ModelUtils.ComputeCentroid(mItems, dataset, type)); // throws ArgumentValueException
}
public SparseVector<double> ComputeCentroid(IUnlabeledExampleCollection<SparseVector<double>> dataset, CentroidType type)
{
return ModelUtils.ComputeCentroid(mItems, dataset, type); // throws ArgumentNullException, ArgumentValueException, InvalidOperationException
}
// *** Computation of centroids ***
public static SparseVector <double> ComputeCentroid(IEnumerable <SparseVector <double> > vecList, CentroidType type)
{
Utils.ThrowException(vecList == null ? new ArgumentNullException("vecList") : null);
Dictionary <int, double> tmp = new Dictionary <int, double>();
Dictionary <int, int> c = new Dictionary <int, int>();
int vecCount = 0;
foreach (SparseVector <double> vec in vecList)
{
foreach (IdxDat <double> item in vec)
{
if (tmp.ContainsKey(item.Idx))
{
tmp[item.Idx] += item.Dat;
c[item.Idx] += 1;
}
else
{
tmp.Add(item.Idx, item.Dat);
c.Add(item.Idx, 1);
}
}
vecCount++;
}
if (vecCount == 0)
{
return(new SparseVector <double>());
}
SparseVector <double> centroid = new SparseVector <double>();
switch (type)
{
case CentroidType.Sum:
foreach (KeyValuePair <int, double> item in tmp)
{
centroid.InnerIdx.Add(item.Key);
centroid.InnerDat.Add(item.Value);
}
break;
case CentroidType.Avg:
foreach (KeyValuePair <int, double> item in tmp)
{
centroid.InnerIdx.Add(item.Key);
centroid.InnerDat.Add(item.Value / (double)vecCount);
}
break;
case CentroidType.AvgNonZero:
foreach (KeyValuePair <int, double> item in tmp)
{
centroid.InnerIdx.Add(item.Key);
centroid.InnerDat.Add(item.Value / (double)c[item.Key]);
}
break;
case CentroidType.NrmL2:
double vecLen = 0;
foreach (KeyValuePair <int, double> item in tmp)
{
vecLen += item.Value * item.Value;
}
vecLen = Math.Sqrt(vecLen);
if (vecLen > 0)
{
foreach (KeyValuePair <int, double> item in tmp)
{
centroid.InnerIdx.Add(item.Key);
centroid.InnerDat.Add(item.Value / vecLen);
}
}
break;
}
centroid.Sort();
return(centroid);
}
public static SparseVector <double> ComputeCentroid(IEnumerable <SparseVector <double> .ReadOnly> vec_list, CentroidType type)
{
Utils.ThrowException(vec_list == null ? new ArgumentNullException("vec_list") : null);
Dictionary <int, double> tmp = new Dictionary <int, double>();
int vec_count = 0;
foreach (SparseVector <double> .ReadOnly vec in vec_list)
{
foreach (IdxDat <double> item in vec)
{
if (tmp.ContainsKey(item.Idx))
{
tmp[item.Idx] += item.Dat;
}
else
{
tmp.Add(item.Idx, item.Dat);
}
}
vec_count++;
}
Utils.ThrowException(vec_count == 0 ? new ArgumentValueException("vec_list") : null);
SparseVector <double> centroid = new SparseVector <double>();
switch (type)
{
case CentroidType.Sum:
foreach (KeyValuePair <int, double> item in tmp)
{
centroid.InnerIdx.Add(item.Key);
centroid.InnerDat.Add(item.Value);
}
break;
case CentroidType.Avg:
foreach (KeyValuePair <int, double> item in tmp)
{
centroid.InnerIdx.Add(item.Key);
centroid.InnerDat.Add(item.Value / (double)vec_count);
}
break;
case CentroidType.NrmL2:
double vec_len = 0;
foreach (KeyValuePair <int, double> item in tmp)
{
vec_len += item.Value * item.Value;
}
Utils.ThrowException(vec_len == 0 ? new InvalidOperationException() : null);
vec_len = Math.Sqrt(vec_len);
foreach (KeyValuePair <int, double> item in tmp)
{
centroid.InnerIdx.Add(item.Key);
centroid.InnerDat.Add(item.Value / vec_len);
}
break;
}
centroid.Sort();
return(centroid);
}
