3

am I right in the assumption that both linear and logistic regression algorithms can be represented as the simplest form of neural networks,a perceptron, which consists of a two layers, an Input and an output, that combines all the input each with it's own respective weight to produces a prediction or classification polynomial?

if this is the case, is anything achievable by Linear or Logistic Regression is also achievable by a perceptron?

Raed Tabani
  • 131
  • 1

1 Answers1

3

Yes, logistic regression can be viewed as a simple neural network with multiple inputs, a single neuron, a single output, and using the logistic function as the activation function.

No, linear regression is not a form of a neural network (a neural network would have a activation function; but there's nothing like that present in linear regression).

D.W.
  • 167,959
  • 22
  • 232
  • 500