IS5740: Management Support Systems

1
IS5740 Management Support Systems

• Neural Computing Applications and Genetic
  Algorithms(Source Chapters 17 and 18, Turban
  and Aronson)

2
Introduction

• ANN Concepts and Fundamentals
• Benefits and limitations of ANN
• Representative Business Applications of ANN
• Integration of ANN and ES
• Genetic Algorithms Concepts
• Representative Business Applications of GA

3
ANN Concepts

• Neural Computing - Constructing computers that
  mimic certain processing capabilities of the
  human brain
• Knowledge representations based on    Massive
  parallel processing   Fast retrieval of large
  amounts of information    The ability to
  recognize patterns based on historical cases
• Neural Computing Artificial Neural Networks
  (ANNs) 

4
The Biology Analogy Biological Neural Networks

• Neurons Brain Cells  Nucleus (at the
  center)  Dendrites provide inputs   Axons
  send outputs
• Synapses increase or decrease connection strength
  and causes excitation or inhibition of subsequent
  neurons 

5
Biological and Artificial Neural Networks
6
Artificial Neural Networks (ANN)

• A model that emulates a biological neural network
• Software simulations of the massively parallel
  processes that involve processing elements
  interconnected in a network architecture
• Originally proposed as a model of the human
  brains activities
• The human brain is much more complex  

7
Neural Network Fundamentals

• Components and Structure    Processing
  Elements    Network    Structure of the
  NetworkProcessing Information in the
  Network    Inputs    Outputs   
  Weights    Summation Function BackPack
  Network Neural System (Electric Load
  Prediction)http//www.zsolutions.com/software.htm
   

8
Benefits of Neural Networks

• Usefulness for pattern recognition, learning,
  classification, generalization and abstraction,
  and the interpretation of incomplete and noisy
  inputs
• Specifically - character, speech and visual
  recognition
• Potential to provide some of human problem
  solving characteristics
• Ability to tackle new kinds of problems
• Robustness
• Fast processing speed
• Flexibility and ease of maintenance
• Powerful hybrid systems

9
Limitations of Neural Networks

• Do not do well at tasks that are not done well by
  people
• Lack explanation capabilities
• Limitations and expense of hardware technology
  restrict most applications to software
  simulations
• Training times can be excessive and tedious
• Usually requires large amounts of training and
  test data 

10
Representative Business ANN Applications

• Accounting
• Identify tax fraud
• Enhance auditing by finding irregularities
• Finance
• Signatures and bank note verifications
• Bankruptcy prediction
• Customer credit scoring
• Credit card approval and fraud detection
• Pricing initial public offerings
• Human Resources
• Predicting employees performance and behavior
• Determining personnel resource requirements

11
Representative Business ANN Applications

• Marketing
• Consumer spending pattern classification
• New product analysis
• Operations Airline Crew Scheduling
• Predicting airline seat demand
• Vehicle routing
• Others
• Assembly and packaged goods inspection
• Fruit and fish grading
• Matching jobs to candidates
• Production/job scheduling

12
Using ANNs for Credit Approval

• Data from the application entered into a separate
  database (Figure 18.2)
• Applications preprocessed manually
• Neural network trained with many good and bad
  risk cases 

13
Neural Network Credit Authorizer

• Construction ProcessStep 1 Collect dataStep 2
  Separate data into training and test setsStep 3
  Transform data into network inputsStep 4
  Select, train and test networkStep 5 Deploy
  developed network application 
• increased loan processor productivity by 25 to 35
  over other computerized tools
• detects credit card fraud

14
Using ANNs for Bankruptcy Prediction

• Application DesignFive Input NodesX1 Working
  capital/total assets X2 Retained
  earnings/total assetsX3 Earnings before
  interest and taxes/total assetsX4 Market value
  of equity/total debtX5 Sales/total assets

15
Using ANNs for Bankruptcy Prediction

• Single Output Node Final classification for each
  firm     Bankruptcy or     Non-bankruptcy
• Development Tool NeuroShell    Three-layer
  network with backpropagation (Fig 18.5)   
  Continuous valued input    Single output node
  0 bankrupt, 1 not bankrupt

16
Using ANNs for Bankruptcy Prediction

• Training    Data Set 129 firms    Training
  Set 74 firms 38 bankrupt, 36 not
• Testing- Test data set 27 bankrupt firms, 28
  non-bankrupt firms- The neural network correctly
  predicted   81.5 percent bankrupt cases   
  82.1 percent non-bankrupt cases 

17
Stock Market Prediction System With Modular
Neural Networks

• Accurate Stock Market Prediction - Complex
  Problem
• Several Mathematical Models - Disappointing
  Results
• Fujitsu and Nikko Securities TOPIX Buying and
  Selling Prediction System
• Input Several technical and economic indexes
• Output Buy/sell timing

18
Stock Market Prediction System (contd.)

• Network ArchitectureNetwork Model (Figure 18.6)
  Training Data     Data Selection    Training
  Data
• Preprocessing Input Indexes - Converted into
  spatial patterns, preprocessed to regularize them
• Moving Simulation Prediction Method (Figure
  18.7) 

19
Examples of Integrated ANNs and Expert Systems

•  1 . Resource Requirements Advisor    Advises
  users on database systems resource
  requirements    Predict the time and effort to
  finish a database project    ES shell AUBREY
  and neural network tool NeuroShell     ES
  supported data collection    ANN used for data
  evaluation    ES final analysis 

20
Examples of

• 2. Personnel Resource Requirements Advisor   
  Project personnel resource requirements for
  maintaining networks or workstations at NASA   
  Rule-based ES determines the final resource
  projections    ANN provides project completion
  times for services requested (Figure 18.9) 

21
Examples of

• 3. Diagnostic System for Singapore Airlines    
  Assist technicians in diagnosing avionics
  equipment    INSIDE (Inertial Navigation System
  Interactive Diagnostic Expert)     Designed to
  reduce the diagnostic time (Figure 18.10)  

22
Genetic Algorithms

• Goal (evolutionary algorithms) Demonstrate
  Self-organization and Adaptation by Exposure to
  the Environment
• System learns to adapt to changes
• Process (Figure 18.11)

23
Definition and Process

• Genetic algorithm "an iterative procedure
  maintaining a population of structures that are
  candidate solutions to specific domain challenges
  (Grefenstette 1982)
• Each candidate solution is called a chromosome
• Chromosomes can copy themselves, mate, mutate
• Use specific genetic operators - reproduction,
  crossover and mutation. 

24
Some Areas of Genetic Algorithm Applications

• Dynamic process control
• Complex design of engineering structures
• Pattern recognition
• Scheduling
• Transportation
• Layout and circuit design

25
Some Business Applications

• Channel 4 Television (England) to schedule
  commercials
• Driver scheduling in a public transportation
  systemJobshop scheduling
• Assignment of destinations to sources
• Trading stocks
• Productivity in whisky making is increased  

26
Representative Commercial Packages

• Evolver (Excel spreadsheet addin)
• OOGA (object-oriented GA for industrial use)
• XperRule Genasys (ES shell with an embedded
  genetic algorithm) 

27
Summary

• ANN can be applied to several difficult problems
  in finance (credit authorization, stock market
  predictions)
• ANN can help interpret information in large
  databases
• Genetic algorithms can be used to solve complex
  optimization problems
• Results improve as knowledge accumulates