Neural Network Architectures

This repository is developed to explain the fundamentals of neural network , how they are better than classic ML algorithms (in several way) and how to develop models using basic NN architecture on different basic data sets.

Data sets :-

• Boston data set - Regression problem .
• IMDB data set - Binary class classification (textual data set) .
• MNIST data set - Multi class classification .
  (description about the data is given in the notebooks)

Prerequisites

• Basic understanding of python, multi layer perceptron , keras , logistic regression , overfitting etc.

The accuracy and the results that are produced are feasible , you can also try to design different architectures using different parameters but keep checking output for overfitting results.

:)