trying to compare neural network with no hidden layers with logistic regression

Mathematically neural network with no hidden layers should be same as logistic regression. On a simple data X class labels Y I am calling patternnet[] and fitglm. The training data provided to train both models is same. When using both models on same test data different accuracy results are obtained. Unable to understand why even the training accuracy of both models is different.

[KAE]

Can you provide a simple code example from one of the built in datasets in Matlab?

Answer 1

[Greg Heath]

No hidden nodes default is the same as

 LINEAR REGRESSION

   NOT

 LOGISTIC REGRESSION

For logistic regression, you have to replace the default linear output layer with a logistic layer.

Hope this helps.

Thank you for formally accepting my answer

Greg