Hands-On Machine Learning with C++

MDS

A multidimensional scaling algorithm is implemented in the Shogun library in the MultidimensionalScaling class. Objects of this class should be configured, along with the number of desired features, with the set_target_dim() method. Then, the fit() method should be used for training. Unlike the previous types, this class provides the transform() method, which transforms the whole dataset into a new number of dimensions. It returns a pointer to the CDenseFeatures type object:

 void MDSReduction(Some<CDenseFeatures<DataType>> features,                   const int target_dim) {   auto IsoMap = some<CMultidimensionalScaling>();   IsoMap->set_target_dim(target_dim);   IsoMap->fit(features);    auto new_features = static_cast<CDenseFeatures<DataType> *>(IsoMap->transform ...