Back to Glossary

Perceptron

Models & Architectures

The simplest neural network with a linear decision boundary.


The Perceptron is a single artificial neuron that computes a weighted sum and classifies via a threshold function.

  • Characteristics: Can only solve linearly separable problems.
  • Training: Perceptron learning rule; converges if data are linearly separable.
  • Significance: Historical foundation for modern neural networks.