Deep Learning

Multi Layer Perceptron ( Core Component of Deep Learning )

Classic Feed Forward Artificial Neural Network
Connected Series of Node ( Directed Acyclic Graph ) No Node is Revisited
Each Node Represents a Function or a Model

When to Use ?

Categorical or Continuous Target Variable
Very Complex Relationships in Data
When Control over Training is very Important ( Lot of Hyperparameters is to be Tuned )

Not to Use

Image Recognition
Time Series