Introduction to Neural Networks and Its Applications

Download Report

Transcript Introduction to Neural Networks and Its Applications

III. Introduction to Neural
Networks And Their
Applications - Basics
Introduction to Neural Networks and Its
Applications
I. Introduction of Neural Networks
II. Application of Neural Networks
III. Theory of Neural Networks
IV. An Example
. Weather Forcasting
I. Introduction of Neural Networks
• Learning in Human Brain
– Neurons
– Connection Between Neurons
• Neural Networks As Simulator For Human
Brain
– Processing Elements or Nodes
– Weights
Main Applications of Neural Networks
• Prediction of Outcomes
• Patterns Detection in Data
• Classification
Why use Neural Networks in Predict
II. Applications of Neural Networks
• Computer Vision : Character recognition
• HNC : Read amount in checks
• NESTOR[Reilly et al , 1990]:Mortgage insurance
decisions
• DAS/LARS[Casselman and Acks,1990] : large
diagnostic system
• DECtalk[Sejnowski and Rosenberg, 1987] : Convert
language to text
• Manufacturing System Controller[Park & Kim, 1991]
: Ford motor Co..
• Investment Decision Making System: Tong Yang
Future & Options in Chicago
III. Theory of Neural Networks
• Network Structure : Layers, Nodes and Weights
Input Layer
Hidden Layer
Output Layer
Training A Neural Networks
• The Key to the success of Neural Networks
use is collecting a lot of good data
• Neural Networks learn from data
• Learning is finding best weights values that
represent the input and output relationship
in Neural Networks
Terms in Neural Networks
Testing and Validating a Neural Networks
• Testing data set : use another new data
• Check the performance of trained Neural
Networks with a testing data
• If it’s performance of test is good , then
check validity of Neural Networks with
another new set of historical data
Prediction with New Data
• If the Neural Network's performance in test
and validation is good , it can be used to
predict outcome of new unseen data
• If the performance with test and validation
is not good, you should collect more data,
add more input variables
IV. A Neural Networks Demo
Intro to neural networks
• http://www.youtube.com/watch?v=DG5UyRBQD4&feature=rellist&playnext=1
&list=PL4FA5D71B0BA92C1C
• Demo for stock market prediction
http://www.youtube.com/watch?v=QoGUE
xdnSDA