Question

Explain the use of neural networking modeling in predictive analytics. Discuss a real-life example of where a neural network model could be used or is used currently in your organization. How would you use SAS Enterprise Miner to create a neural network model of your example?

Answer 1

1.A Neural Network is an artificial system that can perform Intelligent taskssimilar to those performed by human brains.

Neural Networks modelling uses Predictive analytics where it is able to produce correct Output based on the Input

by using historical data which can be usedto produce output when the output is unknown.

Basically in Predictive analytics each neuron takes into acount a set of input values which is linked to individual

weights. The weights are basically Numerical value and contains a value called bias.

Neural networks produce the deisred output based on its weight and bias.

One of thebreal life example of Neural Network is Voice Detection Speech

It is a technology that identifies a speaker and authenticates that the voice belong to the particular persons who

is speaking or not. This technology works by recording the voice of a particular person and digitizing it to create

a unique sample of the voice which can be esily identified .Easc word is broken up into segments comprising

several tones which are agin digitized and captured to create a persons unique voice sample.

2. Usage of SAS enterprise miner to to create a neural network model is a s follows:

SAS Enterprise Miner worksin three units of Neyral Network:

The first one in Input layer whhich ocntains Input variables

The second one is Hidden Layer which makes neural networks moore powerful

The third is the Output Unit which compares the predicted value with the target variables

The SAS Enterprise Miner Implements Toolsfor modelling and utilizing neural network.

The SAS Enterprise Node has two Nodes The Neural Node and Auto Neural Node . The Neural Network Node

train the neural network configuration whereas the Auto Neural Node saerches and finds the best possibe

relationship in the data set and simultaneously train the model based on the data set.