Neural networks

294 views

Published on

Published in: Education
0 Comments
0 Likes
Statistics
Notes
  • Be the first to comment

  • Be the first to like this

No Downloads
Views
Total views
294
On SlideShare
0
From Embeds
0
Number of Embeds
1
Actions
Shares
0
Downloads
7
Comments
0
Likes
0
Embeds 0
No embeds

No notes for slide

Neural networks

  1. 1. DATA WARE HOUSING ANDDATA MINING NEURAL NETWORKS
  2. 2. Contents Neural Networks NN Node NN Activation Functions NN Learning NN Advantages NN Disadvantages
  3. 3. Neural Networks3  Based on observed functioning of human brain.  (Artificial Neural Networks (ANN)  Our view of neural networks is very simplistic.  We view a neural network (NN) from a graphical viewpoint.  Alternatively, a NN may be viewed from the perspective of matrices.  Used in pattern recognition, speech recognition, computer vision, and classification. © Prentice Hall
  4. 4. Neural Networks4  Neural Network (NN) is a directed graph F=<V,A> with vertices V={1,2,…,n} and arcs A={<i,j>|1<=i,j<=n}, with the following restrictions: V is partitioned into a set of input nodes, VI, hidden nodes, VH, and output nodes, VO.  The vertices are also partitioned into layers  Any arc <i,j> must have node i in layer h-1 and node j in layer h.  Arc <i,j> is labeled with a numeric value wij.  Node i is labeled with a function fi. © Prentice Hall
  5. 5. Neural Network Example5 © Prentice Hall
  6. 6. NN Activation Functions6  Functions associated with nodes in graph.  Output may be in range [-1,1] or [0,1] © Prentice Hall
  7. 7. NN Activation Functions7 © Prentice Hall
  8. 8. NN Learning8  Propagate input values through graph.  Compare output to desired output.  Adjust weights in graph accordingly. © Prentice Hall
  9. 9. Neural Networks9  A Neural Network Model is a computational model consisting of three parts:  Neural Network graph  Learning algorithm that indicates how learning takes place.  Recall techniques that determine hew information is obtained from the network.  We will look at propagation as the recall technique. © Prentice Hall
  10. 10. NN Advantages10  Learning  Can continue learning even after training set has been applied.  Easy parallelization  Solves many problems © Prentice Hall
  11. 11. NN Disadvantages11  Difficult to understand  May suffer from overfitting  Structure of graph must be determined a priori.  Input values must be numeric.  Verification difficult. © Prentice Hall
  12. 12. Applications of Neural12 Networks  Prediction – weather, stocks, disease  Classification – financial risk assessment, image processing  Data association – Text Recognition (OCR)  Data conceptualization – Customer purchasing habits  Filtering – Normalizing telephone signals (static)

×