Machine Learning and Data Mining in Pattern Recognition

Name: Machine Learning and Data Mining in Pattern Recognition | 8th International Conference, MLDM 2012, Berlin, Germany, July 13-20, 2012, Proceedings
Brand: Springer
Price: 53.49 EUR
Availability: OnlineOnly

8th International Conference, MLDM 2012, Berlin, Germany, July 13-20, 2012, Proceedings

Petra Perner(Herausgeber*in)

Springer (Verlag)

1. Auflage

Erschienen am 2. Juli 2012

XIII, 680 Seiten

E-Book

PDF mit Wasserzeichen-DRM

Systemvoraussetzungen

978-3-642-31537-4 (ISBN)

53,49 €inkl. 7% MwSt.

Systemvoraussetzungen

für PDF mit Wasserzeichen-DRM

E-Book Einzellizenz

Als Download verfügbar

Beschreibung

Weitere Details

Weitere Ausgaben

Inhalt

Title
Preface
Table of Contents
Theory
Bayesian Approach to the Concept Drift in the Pattern Recognition Problems
Introduction
Bayesian Approach to the Problem of Concept Drift for the Pattern Recognition Problem
Dynamic Programming Procedure for the Estimation of the Decision Rule Parameters under Concept Drift
Experimental Evaluation
``Ground-Truth'' Experiments
Case Study: ``Spam'' E-Mail Problem
Conclusion
References
Transductive Relational Classification in the Co-training Paradigm
Introduction
Related Work
Transductive Learning
Multi-view Learning
Co-training
The CoTReC Method
Constructing Multi-views from Relational Data
Reliability Measures
Computing Reliability Thresholds
Implementation Considerations
Experiments
Datasets
Experimental Results
Conclusions
References
Generalized Nonlinear Classification Model Based on Cross-Oriented Choquet Integral
Introduction
Signed Efficiency Measure, Choquet Integral, and Classification
Signed Efficiency Measure
Choquet Integral
Classification by Choquet Integral Projections
Generalized Classification Model by Cross-Oriented Projection
Algorithms
Experiments
Simulation on Artificial Data
Simulation on Real Data
Conclusions
References
A General Lp-norm Support Vector Machine via Mixed 0-1 Programming
Introduction
A General Lp-norm Support Vector Machine
Optimizing General Lp-norm Support Vector Machine
Experiment
Experimental Settings
Experimental Results
Conclusions
References
Reduction of Distance Computations in Selection of Pivot Elements for Balanced GHT Structure
Introduction
Distance Matrices
Selection of Pivot Elements
Improvements of the Local Optimization Method
Interval Model of Distance Calculations
Conclusions
References
Hot Deck Methods for Imputing Missing Data
Introduction
Review of Literature
Simulation Design
Research Questions
Factorial Design
Quality Criteria
Simulation Details
Results
Donor Limitation Impact
Analysis of Donor Limit Influencing Factors
Analysis of Donor Limitation Advantages
Conclusions
References
BINER
Introduction
Motivation and Contribution
Organization of Paper
Problem Formulation
Related Work
Intuition and Methodology of BINER
The BINER Algorithm
Complexity Analysis
Experimental Study
Performance Model
Results
Discussion
Conclusions
References
A New Approach for Association Rule Mining and Bi-clustering Using Formal Concept Analysis
Introduction
Terminology
FIST Algorithm
Database Preprocessing
Mining Frequent Closed Patterns
Generating Conceptual Clusters and Bases of Association Rules
Performance Analysis
Conclusions
References
Top-N Minimization Approach for Indicative Correlation Change Mining
Introduction
Preliminaries
Correlation Based on k-way Mutual Information
Problem of Mining Top-N Correlation Contrast Sets
Extracting Top-N Correlation Contrast Sets with Extended Double-Clique Search
Double-Clique Search with Dynamic Update on Base Graph
Pruning Mechanism
Algorithm for Extracting Top-N Correlation Contrast Sets
Experimental Results
Computational Performance
Extracted Correlation Contrast Sets
Concluding Remark
References
Theory: Evaluation of Models and Performance Evaluation Methods
Selecting Classification Algorithms with Active Testing
Background and Motivation
Relative Landmarks
Active Testing
Empirical Evaluation
Evaluation Methodology and Experimental Set-up
Results
Related Work in Other Scientific Areas
Significance and Impact
References
Comparing Logistic Regression, Neural Networks, C5.0 and M5´ Classification Techniques
Introduction
The Proposed Algorithm
Experimental Setup
Neural Networks
Logistic Regression
C5.0 and M5´
Data
Experiments
Experimental Results
Discussion and Conclusions
Future Work
References
Unsupervised Grammar Inference Using the Minimum Description Length Principle
Introduction
Related Work
Minimum Description Length Approach
Grammar Description Length (GDL)
Derivation Description Length (DDL)
Operators
Experimental Result and Evaluation
Conclusion and Future Work
References
How Many Trees in a Random Forest?
Introduction
Related Work
Random Trees and Random Forests
Density-Based Metrics for Datasets
Datasets
Experimental Methodology
Results and Discussion
Conclusion
References
Theory: Learning
Constructing Target Concept in Multiple Instance Learning Using Maximum Partial Entropy
Introduction
Partial Entropy
Approaching MIL by Maximizing Partial Entropy
Implementation and Experiments
Artificial Datasets
MUSK Datasets
Conclusion
References
A New Learning Structure Heuristic of Bayesian Networks from Data
Introduction
BNs and Structure Learning from Data Problems
A New Clustering-Based Heuristic: Theoretical Framework and Methodology
Theoretical Background
Procedure and Applied Methodologies
Experimentations Procedures
Data-Bases
Clustering
The Classical Learning Structure Compared to Our New Heuristic
Both Attained Structures' Relevant Inferences and Result Comparisons
Discussion
Conclusion
References
Discriminant Subspace Learning Based on Support Vectors Machines
Introduction
Principal Component Analysis (PCA)
Linear Discriminant Analysis (LDA)
Margin Maximizing Discriminant Analysis (MMDA)
Proposed Approach
Definition and Derivation of the Problem
Deflation Procedure
Within-Class Variance Deflation
Deflated Within-Class Support Vector Machines
Deflation Kernel
Feature Extraction
Experimental Results
Use of WSVDA for Visualization Purposes
Dimensionality Reduction and Classification Results
Conclusions
References
A New Learning Strategy of General BAMs
Introduction
Bidirectional Associative Memory Models
Related Works
Studies about BAMs
Learning Process of BAMs
Given Learning Strategies
Our Approach
Our Learning Strategy
Advantages of the New Learning Strategy
Experiments
Experiment 1 : Our Model Compared to Non-iterative Learning Models
Experiment 2 : Our Model Compared to Iterative Learning Models
Conclusion
References
Proximity-Graph Instance-Based Learning, Support Vector Machines, and High Dimensionality: An Empirical Comparison
Introduction
Proximity Graphs
Reducing the Size of the Training Data
Experiments and Results
Concluding Remarks
References
Theory: Clustering
Semi Supervised Clustering: A Pareto Approach
Introduction
Related Works
Clustering
Semi Supervised Clustering
Problem Definition
Clustering Algorithm
Centroid Update and Assignment
Multi Objective Evolutionary Algorithms
Experiments
Data Sets
Clustering Evaluation Measure
Analysis
Conclusion
References
Semi-supervised Clustering: A Case Study
Introduction
Semi-supervised Clustering
Unsupervised Clustering with K-Means
Unsupervised Clustering with Expectation Maximization
Semi-supervised Clustering by Seeding
Semi-supervised Clustering with Pairwise Constraints
Case Study
Data Description
Experimental Results
Discussion
Conclusions
References
SOStream: Self Organizing Density-Based Clustering over Data Stream
Introduction
Related Work
SOStream Framework
SOStream Overview
Density-Based Centroid
SOStream Algorithm
Online Merging
Experiments
Synthetic Data
Real-World Dataset
Parameter Analysis
Scalability and Complexity of SOStream
Conclusion
References
Clustering Data Stream by a Sub-window Approach Using DCA
Introduction
Clustering Data Stream Based on Sub-windows
The First Analysis: Comparison between Independent Local Clustering and Real Clustering
The Second Analysis: The Adequation between Independent Local Clustering and Global Clustering with the Same Clustering Algorithm
A DCA Clustering Algorithm
General DC Programs
Difference of Convex Functions Algorithms (DCA)
DCA for Solving the Clustering Problem via MSSC (Minimum Sum of Squares Clustering) Formulation
Numerical Experiments
Conclusion
References
Improvement of K-means Clustering Using Patents Metadata
Introduction
Patent Analysis
Prior Work
K-means Algorithm
Data Model for Patents
Weighting Functions
Implementation and Evaluation
Conclusions
References
WebMining
Content Independent Metadata Production as a Machine Learning Problem
Introduction
Data Acquisition
Classifiers Predictions for Metadata Production
Experimental Protocols
Results and Discussions
Association Rules Generation for Metadata Production
Experimental Protocol
Initial Results
Rules Pruning
Improving FP-Growth with Efficient 2 Based Rules Pruning
Results and Discussion
Conclusion
References
Discovering K Web User Groups with Specific Aspect Interests
Introduction
Related Works
Model
Problem Formulation
Model Overview
Generalized Mallows Model(GMM) over Permutations
Generative Process
Analysis of Parameters Setting
Inference
Experiment
Datasets
Evaluation Methodology
Parameter Setting
Comparison of Grouping Performance with K-means Baseline
Illustration of User Groups with Specific Aspect Interests
Conclusion
References
Image Mining
An Algorithm for the Automatic Estimation of Image Orientation
Introduction
The Approach for Automatic Image Orientation Estimation
Experimental Results
Conclusions
References
Multi-label Image Annotation Based on Neighbor Pair Correlation Chain
Introduction
Model Label Correlation under a Graph
Multi-label Image Annotation Based on Neighbor Pair Correlation
Experiments
Conclusions
References
Enhancing Image Retrieval by an Exploration-Exploitation Approach
Introduction
Nearest-Neighbor Relevance Feedback for Relevance Score
The Exploration Phase
Experimental Results
Datasets
Experimental Setup
Results
Conclusion
References
Finding Correlations between 3-D Surfaces: A Study in Asymmetric Incremental Sheet Forming
Introduction
Previous Work
Grid Representation
Springback Measurement
Surface Representation (The Local Geometry Matrix)
Classifier Generation
Classifier Application
Evaluation
Conclusions and Perspectives
References
Data Mining in Biometry and Security
Combination of Physiological and Behavioral Biometric for Human Identification
Introduction
Background
Multimodal Identification Scheme
PCA- LDA
MLP
J48 Classifier
Naïve Bayes Classifier
SMO Classifier
Experimental Results and Discussion
Conclusions and Further Plan
References
Detecting Actions by Integrating Sequential Symbolic and Sub-symbolic Information in Human Activity Recognition
Introduction
Methods
Dataset
Results
Object Recognition
Action Recognition
Conclusions
References
Computer Recognition of Facial Expressions of Emotion
Introduction
Emotion Recognition Problem
Facial Detection
Facial Expression Representation
Emotion Recognition
Experimental Results
Conclusion
References
Data Mining in Medicine
Outcome Prediction for Patients with Severe Traumatic Brain Injury Using Permutation Entropy Analysis of Electronic Vital Signs Data
Introduction
Method
Ordinal Pattern and Permutation Entropy
Multivariate Time Series
Evaluations
Experiments and Results
Data and Setup
Prediction for Mortality and 3-Month GOSE
Baseline
Conclusion
Summary
Future Work
Clinical Implication
References
EEG Signals Classification Using a Hybrid Method Based on Negative Selection and Particle Swarm Optimization
Introduction
Artificial Immune Systems
Particle Swarm optimization
Materials and Methods
EEG Data
Discrete Wavelet Transform: Feature Extraction
Adaptive Particle Swarm Negative Selection: EEG Classification
Experimental Results
Performance Measures
Results and Discussion
Conclusion
References
Data Mining in Environment and Water Quality Detection
DAGSVM vs. DAGKNN: An Experimental Case Study with Benthic Macroinvertebrate Dataset
Introduction
Methods
Support Vector Machine
DAGSVM and DAGKNN
Experimental Tests
Data Description and Test Arrangements
Results
Discussion
References
Image Mining in Medicine
Lung Nodules Classification in CT Images Using Shannon and Simpson Diversity Indices and SVM
Introduction
Methodology
Image Acquisition
Preprocessing with Histogram Equalization
Lung Nodule Segmentation
Features Extraction Using Diversity Indices
Support Vector Machine
Selection of Features Using Stepwise Linear Discriminant Analysis
Validation of the Classification Method
Experiments
Results
Conclusions
References
Comparative Analysis of Feature Selection Methods for Blood Cell Recognition in Leukemia
Introduction
Problem Statement
Feature Selection
Fisher Method
Correlation of the Feature with the Class
The Feature Selection Based on the Application of the Multiple Input Linear SVM
The Feature Selection Based on the Application of Nonlinear Kernel
Principal Component Analysis for Feature Extraction and Selection
Independent Component Analysis for Feature Extraction and Selection
Selection of the Features Using Genetic Algorithm
Ensemble of Classifiers and Final Results of Classification
Conclusions
References
Classification of Breast Tissues in Mammographic Images in Mass and Non-mass Using McIntosh's Diversity Index and SVM
Introduction
Related Work
Materials and Methods
McIntosh's Diversity Index
Histogram
GLCM - Gray Level Co-occurrence Matrix
GLRLM - Gray Level Run Length Matrix
Proposed Methodology
Database
Features Extraction and Classification
Results and Discussion
Conclusion
References
Text Mining
A Semi-Automated Approach to Building Text Summarisation Classifiers
Introduction
Related Work
Problem Formalisation
Classifier Generation Using SARSET (Semi-Automated Rule Summarisation Extraction Tool)
Phrase Identification and Generation of Phrase Variations (Step 1)
Identification of Questionnaires Covered by Identified Phrases (Step 2)
Rule Generation (Steps 3 and 4)
Continuation of the Process or Exit (Steps 5)
Applying Classification Rules to Unseen Documents
The SAVSNET Application
Evaluation
Conclusion
References
A Pattern Recognition System for Malicious PDF Files Detection
Introduction
An Overview of PDF Technology
A Brief History
PDF Structure
Attacks against PDF Documents
Attack Types
Evasion Techniques
Typical Attack Procedure
Related Works on PDF Security
General PDF Security Research
State-of-the-Art Malicious PDF Detection Tools
A New PDF Detector
Feature Extraction
Classification
Results
Data Collection
Feature Extraction
Choice of the Classifier
Accuracy on the Test Set
Weaknesses
Conclusions
References
Text Categorization Using an Ensemble Classifier Based on a Mean Co-association Matrix
Introduction
Problem Overview
Methodology
Methodology Application
Dataset
Preprocessing
Ensemble Building
Evaluation Metrics
Results
Results Validation (Friedman Test)
Discussion
Conclusions and Future Work
References
A Pattern Discovery Model for Effective Text Mining
Introduction
Related Work
Basic Definitions
Mining Informative Contents
A Pattern Fusion Model
Transferable Belief Model (TBM)
Evidential Mapping
Weight Fusion
Reasoning
Experimental Evaluation
Experimental Dataset
Data Preprocessing and Measures
Baseline Models and Settings
Quality of Extracted Features
PFM vs. Pattern Mining Models
PFM vs. Term-Based Models
Conclusion
References
Investigating Usage of Text Segmentation and Inter-passage Similarities to Improve Text Document Clustering
Introduction
Related Work
Basic Idea
Text Segmentation
Similarity Computation
Traditional Inter-document Similarity
Passage-Based Inter-document Similarity
Combined Similarity Measure
Experimental Results
Data Sets
Evaluation Measure
Clustering Algorithm
Baseline Approach
Results
Conclusion and Future Work
References
Data Mining in Network
Mining Ranking Models from Dynamic Network Data
Introduction
Related Work
Learning Ranking Functions with Different Time Windows
Experiments
Conclusions
References
Machine Learning-Based Classification of Encrypted Internet Traffic
Introduction
Related Work
BitTorrent Architecture Overview
Proposed P2P Identification System
Database
Pre-filtering
Flow Conversion and Features Computation
Feature Selection and Classification
Support Vector Machines
Experimental Results
Conclusion
References
Application of Bagging, Boosting and Stacking to Intrusion Detection
Intrusion Detection System
Data Mining for IDS
Ensemble Classifier
Bagging
Boosting
Stacking
Experimental Settings
Intrusion Dataset
Performance Metric
Experimental Settings
Conclusions
References
Applications of Data Mining in Process Automation, Organisation Change Management, Telecommunication and Post Services
Classification of Elementary Stamp Shapes by Means of Reduced Point Distance Histogram Representation
Introduction
Features Extraction/Reduction
Point Distance Histogram Representation
Features Dimensionality Reduction by Means of PCA/LDA
Classification
Experimental Results
Summary
References
A Multiclassifier Approach for Drill Wear Prediction
Introduction
The Drill Wear Problem
The Drilling Process
Methodology
Feature Selection
Supervised Classification
Classification Algorithms
Validation Method
Multiclassifier Approach
Experimental Results
Burr detection
Roughness Detection
Conclusions
References
Measuring the Dynamic Relatedness between Chinese Entities Orienting to News Corpus
Introduction
Related Work
Methods
Outline
Development Law of Dynamic Relatedness between Entity-Pair
Development Process of Dynamic Relatedness
Experiments
Dataset
Evaluation of Metrics and Methods
Results and Analysis
Conclusions
References
Prediction of Telephone User Attributes Based on Network Neighborhood Information
Introduction
Some Interesting Scenarios
Data Description and Preparation
Exploratory Analysis and Learning Approach
Isolated Link Approach
SMS and Call Metrics
Gender Prediction
Age Prediction
Ego-Network Approach
Network Metrics
Gender Prediction
Age Prediction
Conclusion and Future Work
References
Data Mining in Biology
A Hybrid Approach to Increase the Performanceo f Protein Folding Recognition Using Support Vector Machines
Introduction
Protein Database and Its Features
Protein Database
Features Vectors
Machine Learning Classifiers
Support Vector Machine Classifiers (SVM)
Quadratic Discriminant Analysis
Feature Selection
Generalized Linear Model
Sequential Feature Selection Algorithm
Feature Selection Algorithms
Results
Conclusions and Future Work
References
Author Index

Systemvoraussetzungen

Als PDF speichern Als Link merken