4 ms·

Question: Why combining indicators doesn't work?

• wisty 16y ago
  The curse of dimensionality, I'd guess.
• ggruschow 16y ago
  It can, but it's usually an impossibly huge search for a way to turn garbage into gold.
• pictional 16y ago
  Overfitting, curse of dimensionality, etc. As in many machine learning tasks, extracting good starting-point features is very very important, and that often requires domain-specific knowledge.