EasyDeepLearn

What is feature hashing (the 'hashing trick') and when is it useful?

medium

Answer

  • Hash each category / n-gram to an index in a fixed-size vector (eg,  218  columns)(eg, \;2^{18}\;\mathrm{columns}) using a fast hash function.
  • Handles unbounded / very high cardinality without an explicit vocabulary (users, URLs, streaming categoricals).
  • Trades collisions for a fixed memory footprint — controllable by the number of hash buckets.
  • Used in Vowpal Wabbit, online learning, and large-scale linear models with billions of features.
  • Downside: model interpretability drops; two colliding categories become inseparable.
Check yourself — multiple choice
  • Hash the target instead of features
  • Hash categories/n-grams into a fixed-size vector — bounded memory, tolerates collisions
  • Same as one-hot encoding
  • Only used for computer vision

Hashing trick = fixed-size feature vector with tolerated collisions ⇒ great for very high cardinality.

#encoding#features

Practise Supervised Learning

215 interview questions in this topic.

Related questions