Flash intro to Neural Network

Antony Samuel B

Logistic Regression

  • A logistic regression is used to classify between two outputs A and B or 0 and 1.
  • Here we see a representation of a logistic regression with 3 inputs and a classification type ouput.

plot of chunk plot1

Multiclass logistic regression

  • A multiclass logistic regression is used when we have to classify between multiple outputs.
  • Here we see such a regression's representation with 4 inputs and 3 classification ouputs.

plot of chunk plot2

Neural Network

  • A neural Network can be thought of stacked multiclass logistic regression.
  • The below NN can be thought of as 2 stacked multiclass logistic regression, with the ouputs of the first serving as the inputs to the next.

plot of chunk plot3

Thanks

  • If you liked this intro go through Andrew Ng's course in coursera.org.
  • Thanks.