Next:
List of Figures
Up:
No Title
Previous:
Acknowledgements
Contents
List of Figures
Introduction
Our Task
Reader's Guide
Data Mining
What is Data Mining
Training and Test Sets
Examples from Real Life
Learning
Inferring Information Using Deduction or Induction
Supervised and Unsupervised Learning
Representation of Knowledge
Propositional Representations
First Order Logic
Other Representations
Rule Generation
Semantics of Definite Rules
Non-Monotonic and Default Reasoning
Problems and Challenges in Data Mining
Noisy Data
Difficult Training Set
Databases are Dynamic
Databases may be Huge
Evaluation Criteria for Data Mining Systems
Cost of the Learning Set
Time and Memory Constraints During Learning
Correct Prediction
Prediction when the Training Set is Noisy
Efficiency of Learned Knowledge
Syntactic and Semantic Simplicity of Result
Methods for Data Mining
Bayesian Approach
ID3
Rough Sets Approach
Rough Sets
Information System
Discerning Objects
Discernibility Matrix
Discernibility Functions
Reducing Representation
Upper and Lower Approximation
From Reducts to Rules
Rough Membership Function
An Algorithm for Default Rules Generation
Definitions
A Motivating Example
The Algorithm
Traversal Example
Classification
Methods for Using Existing Rules
Accuracy of the Rules:
Weight of Evidence Measure
Detection of Underlying Patterns
Rule Generation
Determination of Class Membership
Rules Available to the Classifier
Problems and Opportunities
Conflicting Rules
Rules in a Sub-Lattice
Multiple Equal Rules
The Implemented System
Testing or Classification of New Objects
Statistics
Validity Measure
Methods
First Rule
The Rule with Highest Accuracy Matching
Simple Voting
Voting by the Weight of Evidence Measure
Ignoring Blocks
Using Certain Attributes
Using Certain Rule bases
Possibility for Later Additions
Results
State of the Art and Related Work
The Data Mining Process
Specific Data Mining Applications
Rough Sets Applications
Conclusions and Future Work
Summary of Important Features
Conclusion
Improvements to RGEN
Improvements to RCLASS
References
User's Guide
System Overview
Main Program
RGEN - Rules Generation
The User Interface
Running RGEN Directly from the Command Prompt
RCLASS - Classification
The User Interface
Running RCLASS Directly from the Command Prompt
Experimenting with RGEN and RCLASS
Preparing for the Tests
The Experiments
Data From UCI Repository
Data From SCDS
Possible Tests not Performed
File formats
Table File
Dictionary file
Rule File
The RGEN Rule File
Using RSES Rules in RCLASS
Using RGEN Rules in RSES
Implementation Issues
Program Structure Overview
Internal List Format for Rules
Internal Format for the Dictionary
Classes
AVPair:public TObject
AttrMap
AttrMapList
AttrList
Block:public TObject, public ObjList
BlockList:public TObject, public ObjList
Classifier
Rule:public TObject, public ObjList
RuleList:public TObject, public ObjList
TDTable
Changes Made to RGEN
Faster Reduct Calculation
Fixing the Dictionary Format
Making the Rules Look Like in RSES
Traversing the Lattice Without Calculating Rules
Our Code
makefile
AVPair.h
AVPair.C
AttrList.h
AttrList.C
block.h
block.C
blocklst.h
blocklst.C
classifi.h
classifi.C
List.h
List.C
Minimize.h
Minimize.C
rclass.C
rule.h
rule.C
rulelist.C
rulelist.h
Code for the Script rgen2rses
Tk/Tcl code for the Graphical User Interface
Main program - file MAIN
RGEN user interface - file RGEN
RCLASS user interface - file RCLASS
About this document ...
Helge Grenager Solheim
Sat May 4 03:30:02 MET DST 1996