Machine Learning for Developers

Similarities and differences between a perceptron and ADALINE

We have covered a simplified explanation of the precursors of modern neural networks. As you can see, the elements of modern models were almost all laid out during the 1950s and the 1960s! Before continuing, let's try to compare the approaches:

• Similarities:
  • They are both algorithms (it's important to stress that)
  • They are applied to single-layer neural models
  • They are classifiers for binary classification
  • Both have a linear decision boundary
  • Both can learn iteratively, sample by sample (the perceptron naturally, and ADALINE via stochastic gradient descent)
  • Both use a threshold function
• Differences:
  • The perceptron uses the final class decision to train weights
  • ADALINE uses ...