de
en
Schliessen
Detailsuche
Bibliotheken
Projekt
Impressum
Datenschutz
Schliessen
Publizieren
Besondere Sammlungen
Digitalisierungsservice
Hilfe
Impressum
Datenschutz
zum Inhalt
Detailsuche
Schnellsuche:
OK
Ergebnisliste
Titel
Titel
Inhalt
Inhalt
Seite
Seite
Im Werk suchen
Novel methods for mining and learning from data streams / vorgelegt von Ammar Shaker. Paderborn, 2017
Inhalt
Contents
List of Figures
List of Tables
1 Introduction
1.1 Application Example
1.2 Learning from Data Streams
1.3 Incremental, Adaptive and Evolving Learning
1.4 Contribution and Outline of the Thesis
2 Background
2.1 Machine Learning
2.2 Supervised Learning from Data Streams
2.3 Concept Change over Time
2.4 Change Detection Methods
2.5 Adaptive Supervised Learning: Related Work
2.5.1 Rule-based learning
2.5.2 Decision trees learning
2.5.3 Instance-based learning
2.5.4 Ensemble methods
3 Instance-Based Classification and Regression
3.1 Instance-Based Learning
3.1.1 Classification
3.1.2 Regression
3.2 Instance-Based versus Model-Based Learning
3.3 Instance-Based Learning on Data Streams
3.4 IBLStreams
3.4.1 Classification
3.4.2 Regression
3.4.3 Parameter adaptation in IBLStreams
3.4.4 Implementation issues
3.5 Experiments
3.5.1 IBLStreams versus other instance-based methods
3.5.2 Evaluating the parameter adaptation schemes
3.5.3 IBLStreams versus state-of-the-art model-based methods
3.5.3.1 Classification
3.5.3.2 Regression
3.6 Discussion and Conclusion
4 Evolving Fuzzy Pattern Trees
4.1 Introduction to Fuzzy Sets
4.1.1 Operations on Fuzzy Sets
4.1.2 Aggregation Operations on Fuzzy Sets
4.2 Data-Driven Fuzzy Modeling
4.2.1 Fuzzy Subsethood-Based Algorithm
4.2.2 Fuzzy Decision Trees
4.3 Fuzzy Pattern Trees
4.3.1 Bottom-Up Induction of Fuzzy Pattern Trees
4.3.2 Top-Down Induction of Fuzzy Pattern Trees
4.4 Evolving Fuzzy Pattern Trees
4.4.1 Performance Monitoring and Hypothesis Testing
4.4.2 Summary of the Algorithm
4.4.3 Refinements on the Neighbor Trees Generation
4.5 Empirical Evaluation
4.5.1 Performance Comparison
4.5.1.1 Synthetic Data
4.5.1.2 Real Data
4.5.2 Model Size
4.5.3 Sensitivity Towards Significance Levels and Operators Retraining
4.6 Summary and Conclusion
5 Survival Analysis on Event Streams
5.1 Introduction
5.2 Survival Analysis
5.2.1 Censored data
5.2.2 Survival Functions
5.2.3 Estimating the Survival Function
5.2.4 Prognostic Factors for Survival
5.3 Survival Analysis on Data Streams
5.3.1 Left Censoring
5.3.2 Parallel Event Sequences
5.3.3 Adaptive ML Estimation
5.4 Case Study: Earthquake Analysis
5.4.1 Data Generation
5.4.2 Results
5.5 Case Study: Twitter Data
5.6 Conclusion
6 Recovery Analysis for Adaptive Learning
6.1 Introduction
6.2 Learning under concept drift
6.3 Recovery Analysis
6.3.1 Main idea and experimental protocol
6.3.2 Bounding the optimal generalization performance
6.3.3 Recovery measures
6.3.4 Defining pure streams
6.3.5 Further practical issues
6.4 A comparison of algorithms
6.5 Experiments and results
6.5.1 Binary classification
6.5.2 Multiclass classification
6.5.3 Regression
6.5.4 Recovery measures
6.5.5 Summary of the experiments
6.6 Conclusion
7 Conclusion
7.1 Original Contributions
7.2 Future Research
A Methods
A.1 Adaptive Hoeffding Tree
A.2 Adaptive Model Rules
A.3 Fast Incremental Model Trees with Drift Detection
A.4 FLEXible Fuzzy Inference Systems
B MOA
B.1 Stream Generators
B.2 Online Evaluation
C M-Tree
C.1 Distance Function
D Data Sets
D.1 Synthetic Data Sets
D.1.1 Hyperplane data
D.1.2 Distance to hyperplane data
D.1.3 Random trees data
D.1.4 Radial basis function data
D.1.5 SEA concept functions
D.1.6 STAGGER concept functions
D.2 Synthetic Data Manipulation
D.2.1 Concept drift simulation
D.2.2 Sampling drift simulation
D.3 Real Data Sets
D.3.1 Cover type data
D.3.2 Mushroom data
D.3.3 Page blocks data
D.3.4 Letter recognition
D.3.5 StatLog (shuttle) data
D.3.6 Skin segmentation data
D.3.7 MAGIC gamma telescope data
D.3.8 Breast cancer Wisconsin
D.3.9 Parkinson's telemonitoring data
D.3.10 Slice localization data
D.3.11 Bank32h
D.3.12 Census-house
D.4 Event Streams
D.4.1 Earthquake event stream
D.4.2 Twitter stream
E Incremental Statistics
E.1 Incremental Moments
E.2 Shifting Moments
Bibliography
Die detaillierte Suchanfrage erfordert aktiviertes Javascript.