Return to Welcome Screen
Introduction 1. The APS Environment 1.1. Getting Started 1.1.1. Starting APS 1.1.2. Loading Data 1.1.3. Evolving a Model 1.1.4. Optimizing a Model Evolved by APS 1.1.5. Evaluating the Accuracy of a Model 1.1.6. Testing the Generalizing Capabilities of a Model 1.1.7. Generating Code Automatically 1.1.8. Visualizing Parse Trees 1.1.9. Scoring a Model 1.2. APS Tools 1.2.1. Function Finding Tools 1.2.2. Classification Tools 1.2.3. Time Series Analysis Tools 1.3. APS Panels 1.3.1. Report 1.3.2. Data 1.3.2.1. Training Data 1.3.2.2. Testing Data 1.3.3. Settings 1.3.3.1. General Settings 1.3.3.2. Fitness Function 1.3.3.3. Genetic Operators 1.3.3.4. Numerical Constants 1.3.4. Functions 1.3.4.1. Mathematical Functions 1.3.4.2. Static UDFs 1.3.5. Run 1.3.6. History 1.3.7. Results 1.3.7.1. Training Results 1.3.7.2. Testing Results 1.3.8. Predictions 1.3.9. Model 1.4. Evolving APS Models 1.4.1. Preparing and Analyzing Data 1.4.2. Choosing the Function Set 1.4.3. Choosing the Chromosome Architecture 1.4.4. Choosing the Population Size 1.4.5. Choosing the Degree of Genetic Modification 1.4.6. Choosing the Fitness Function 1.4.7. Creating the Model 1.4.8. Evaluating the Model 1.4.9. Modeling from Seed 1.5. Making Predictions with a Model 1.5.1. Scoring from Databases 1.5.2. Scoring from Text Files 1.5.3. Making Predictions with Time Series Models 2. Working with Function Finding Tools 2.1. Loading Data 2.2. Available Functions 2.2.1. Mathematical Functions and Dynamic UDFs 2.2.2. Static UDFs 2.3. Available Chromosomal Architectures 2.4. Fitness Functions 2.5. Available Algorithms 2.6. Evolutionary Strategies 3. Working with Classification Tools 3.1. Loading Data 3.2. Available Functions 3.2.1. Mathematical Functions and Dynamic UDFs 3.2.2. Static UDFs 3.3. Available Chromosomal Architectures 3.4. Fitness Functions and the Rounding Threshold 3.5. Available Algorithms 3.6. Evolutionary Strategies 4. Working with Time Series Prediction Tools 4.1. Loading Data 4.2. Available Functions 4.2.1. Mathematical Functions and Dynamic UDFs 4.2.2. Static UDFs 4.3. Available Chromosomal Architectures 4.4. Fitness Functions 4.5. Available Algorithms 4.6. Evolutionary Strategies 5. The Architectures of APS Learning Algorithms 5.1. Gene Expression Programming 5.1.1. An Overview of Gene Expression Programming 5.1.2. The Architecture of GEP Programs 5.1.3. Multigenic Chromosomes and Linking Functions 5.1.4. User Defined Functions 5.2. Gene Expression Programming with Random Numerical Constants 5.2.1. Function Finding and the Creation of Numerical Constants 5.2.2. The Architecture of GEP-RNC Programs 5.3. Karva Notation: The Language of GEP 5.3.1. Karva Notation and the Manipulation of K-expressions 5.3.2. The Symbols of Karva Notation 6. Genetic Operators 6.1. Mutation 6.2. Inversion 6.3. Transposition 6.3.1. IS Transposition 6.3.2. RIS Transposition 6.3.3. Gene Transposition 6.4. Recombination 6.4.1. One-point Recombination 6.4.2. Two-point Recombination 6.4.3. Gene Recombination 6.5. Genetic Operators for Random Numerical Constants 6.5.1. Dc Mutation 6.5.2. Dc Inversion 6.5.3. Dc IS Transposition 6.5.4. Direct Mutation of Random Numerical Constants 7. Evolution from Existing Models 7.1. Optimizing Models Evolved by APS 7.2. Modifying Models Evolved by APS 7.3. Optimizing an External Model 7.4. Adding a Neutral Gene 8. Choosing the Fitness Function 8.1. Fitness Functions for Function Finding 8.1.1. Relative Error Based Fitness Functions 8.1.1.1. Relative Error with Selection Range 8.1.1.2. Relative/Hits 8.1.2. Absolute Error Based Fitness Functions 8.1.2.1. Absolute Error with Selection Range 8.1.2.2. Absolute/Hits 8.1.3. R-square Based Fitness Function 8.1.4. Mean Squared Error Based Fitness Function 8.1.5. Root Mean Squared Error Based Fitness Function 8.1.6. Mean Absolute Error Based Fitness Function 8.1.7. Relative Squared Error Based Fitness Function 8.1.8. Root Relative Squared Error Based Fitness Function 8.1.9. Relative Absolute Error Based Fitness Function 8.1.10. User Defined Fitness Functions 8.2. Fitness Functions for Classification 8.2.1. Number of Hits 8.2.2. Sensitivity/Specificity Based Fitness Function 8.2.3. PPV/NPV Based Fitness Function 8.2.4. R-square Based Fitness Function 8.2.5. Mean Squared Error Based Fitness Function 8.2.6. Root Mean Squared Error Based Fitness Function 8.2.7. Mean Absolute Error Based Fitness Function 8.2.8. Relative Squared Error Based Fitness Function 8.2.9. Root Relative Squared Error Based Fitness Function 8.2.10. Relative Absolute Error Based Fitness Function 8.2.11. User Defined Fitness Functions 8.3. Fitness Functions for Time Series Prediction 8.3.1. Relative Error Based Fitness Functions 8.3.1.1. Relative Error with Selection Range 8.3.1.2. Relative/Hits 8.3.2. Absolute Error Based Fitness Functions 8.3.2.1. Absolute Error with Selection Range 8.3.2.2. Absolute/Hits 8.3.3. R-square Based Fitness Function 8.3.4. Mean Squared Error Based Fitness Function 8.3.5. Root Mean Squared Error Based Fitness Function 8.3.6. Mean Absolute Error Based Fitness Function 8.3.7. Relative Squared Error Based Fitness Function 8.3.8. Root Relative Squared Error Based Fitness Function 8.3.9. Relative Absolute Error Based Fitness Function 8.3.10. User Defined Fitness Functions 9. Analyzing APS Models Statistically 9.1. Statistical Analysis of Function Finding Models 9.1.1. R-square 9.1.2. Correlation Coefficient 9.1.3. Mean Squared Error 9.1.4. Root Mean Squared Error 9.1.5. Mean Absolute Error 9.1.6. Relative Squared Error 9.1.7. Root Relative Squared Error 9.1.8. Relative Absolute Error 9.2. Statistical Analysis of Classification Models 9.2.1. Classification Error 9.2.2. Classification Accuracy 9.2.3. Confusion Matrix 9.2.4. Sensitivity 9.2.5. Specificity 9.2.6. Positive Predictive Value 9.2.7. Negative Predictive Value 9.2.8. R-square 9.2.9. Correlation Coefficient 9.2.10. Mean Squared Error 9.2.11. Root Mean Squared Error 9.2.12. Mean Absolute Error 9.2.13. Relative Squared Error 9.2.14. Root Relative Squared Error 9.2.15. Relative Absolute Error 9.3. Statistical Analysis of Time Series Prediction Models 9.3.1. R-square 9.3.2. Correlation Coefficient 9.3.3. Mean Squared Error 9.3.4. Root Mean Squared Error 9.3.5. Mean Absolute Error 9.3.6. Relative Squared Error 9.3.7. Root Relative Squared Error 9.3.8. Relative Absolute Error 10. Generating Code Automatically 10.1. Endogenous Code 10.1.1. Karva Code 10.1.2. Expression Trees 10.2. Built-in Grammars 10.2.1. C 10.2.2. C++ 10.2.3. C# 10.2.4. Visual Basic 10.2.5. VB.Net 10.2.6. Java 10.2.7. Java Script 10.2.8. Fortran 10.3. User Defined Grammars 11. Making Predictions with a Model 11.1. Scoring from Databases 11.2. Scoring from Text Files 11.3. Making Predictions with Time Series Models 12. Settings and Features 12.1. Settings and Features for Function Finding Problems 12.1.1. New Run Wizard 12.1.1.1. Run Category Window 12.1.1.2. Data Source Window 12.1.1.3. Text File Window 12.1.1.4. Database Window 12.1.1.5. Training Query Window 12.1.1.6. Testing Query Window 12.1.1.7. Testing Datasets Window 12.1.2. Data Panel 12.1.2.1. Training Set Tab 12.1.2.2. Testing Set Tab 12.1.3. Settings Panel 12.1.3.1. General Settings Tab 12.1.3.2. Fitness Function Tab 12.1.3.3. Genetic Operators Tab 12.1.3.4. Numerical Constants Tab 12.1.4. Functions Panel 12.1.4.1. Functions (Math) Tab 12.1.4.2. Functions (Math) Tab – Edit DDF Window 12.1.4.3. Static UDFs Tab 12.1.4.4. Static UDFs Tab – Edit UDF Window 12.1.5. Run Panel 12.1.6. History Panel 12.1.7. Results Panel 12.1.7.1. Results Panel – General 12.1.7.2. Grid Option 12.1.7.3. Chart Option 12.1.8. Model Panel 12.1.9. Change Seed Window 12.1.10. Scoring 12.1.10.1. Text File Window 12.1.10.2. Database Window 12.2. Settings and Features for Classification Problems 12.2.1. New Run Wizard 12.2.1.1. Run Category Window 12.2.1.2. Data Source Window 12.2.1.3. Text File Window 12.2.1.4. Database Window 12.1.1.5. Training Query Window 12.1.1.6. Testing Query Window 12.1.1.7. Testing Datasets Window 12.2.2. Data Panel 12.2.2.1. Training Set Tab 12.2.2.2. Testing Set Tab 12.2.3. Settings Panel 12.2.3.1. General Settings Tab 12.2.3.2. Fitness Function Tab 12.2.3.3. Genetic Operators Tab 12.2.3.4. Numerical Constants Tab 12.2.4. Functions Panel 12.2.4.1. Functions (Math) Tab 12.2.4.2. Functions (Math) Tab – Edit DDF Window 12.2.4.3. Static UDFs Tab 12.2.4.4. Static UDFs Tab – Edit UDF Window 12.2.5. Run Panel 12.2.6. History Panel 12.2.7. Results Panel 12.2.7.1. Results Panel – General 12.2.7.2. Grid Option 12.2.7.3. Chart Option 12.2.8. Model Panel 12.1.9. Change Seed Window 12.1.10. Scoring 12.1.10.1. Text File Window 12.1.10.2. Database Window 12.3. Settings and Features for Time Series Prediction Problems 12.3.1. New Run Wizard 12.3.1.1. Run Category Window 12.3.1.2. Data Source Window 12.3.1.3. Text File Window 12.3.1.4. Time Series Transformation Window 12.3.1.5. Testing Datasets Window 12.3.2. Data Panel 12.3.2.1. Transformed Time Series Tab 12.3.2.2. Original Time Series Tab 12.3.3. Settings Panel 12.3.3.1. General Settings Tab 12.3.3.2. Fitness Function Tab 12.3.3.3. Genetic Operators Tab 12.3.3.4. Numerical Constants Tab 12.3.4. Functions Panel 12.3.4.1. Functions (Math) Tab 12.3.4.2. Functions (Math) Tab – Edit DDF Window 12.3.4.3. Static UDFs Tab 12.3.4.4. Static UDFs Tab – Edit UDF Window 12.3.5. Run Panel 12.3.6. History Panel 12.3.7. Predictions Panel 12.3.8. Model Panel 12.1.9. Change Seed Window 13. APS Editions 13.1. Editions 13.2. Requirements 13.3. Installation 13.4. Menus 13.4.1. Run 13.4.2. Edit 13.4.3. View 13.4.4. Panels 13.4.5. Scoring 13.4.6. Help 13.5. License 13.6. Demo Learning Resources Support Contacts